By 2025, 25% of Y Combinator’s Winter batch had codebases that were 95% AI-generated. Collins Dictionary named “vibe coding” its 2025 Word of the Year. The AI code generation market hit $4.91 billion.

A UX designer in an academic study said it best: “What used to take me hours, I can now do in two minutes with AI.”

Everyone celebrated the democratisation of software. And rightly so. But here’s the thing nobody’s spending enough time on: the constraints that made software expensive also shaped the experience of using it.

Remove the constraint, and you don’t just get faster software. You get more software. The question is whether more is better - or just more.

The old world: when software was precious

Let me paint you a picture of product management in 2021.

You’re a PM at a fintech. You have 6 engineers. US-based senior developers cost $150–$200 per hour, and when you factor in benefits, recruitment, and overhead, the true cost of a fully employed engineer runs roughly 2.7x their base salary.

Every sprint is a zero-sum game. Every feature you build means something else you don’t.

The ritual goes like this: spend 3 days writing a PRD nobody fully reads. Two design review cycles. A stakeholder alignment meeting. At least one “can we push this to next quarter?” Get engineering estimates that are optimistic by 40%. Build anyway. Ship late. Celebrate like you conquered something.

The process was slow and annoying. But it had a hidden benefit: it was a quality filter.

The pain of getting anything built meant only things worth building got built. Everything had to survive interrogation. ROI calculations. Stakeholder scepticism. Engineering capacity constraints.

Internal tooling? Almost never.

At Wise, a globally scaled fintech where I worked back then, building tools for operations teams was notoriously difficult - not because it was technically hard, but because engineering resources were too scarce to point at anything that didn’t directly move a customer-facing metric.

The ops team needed a custom dashboard? Great idea. Back of the queue behind the 11 customer-facing features committed in Q1.

Software was expensive, so you built less and built it well.

Enter the Vibe coders

In February 2025, Andrej Karpathy - former head of AI at Tesla, co-founder of OpenAI - coined the term “vibe coding.” Describe what you want in plain English. Let the AI write the code. Don’t even look at it (although you should).

A Meta PM named Zevi Arnovitz, who described himself as someone who finds code “terrifying,” told Lenny’s Podcast that AI coding tools felt like being handed “superpowers.” He rebuilt his entire workflow using Cursor and Claude - no engineering background, shipping features independently.

Figma CEO Dylan Field noted that designers, engineers, PMs, and researchers were all starting to “dip their toe into other roles.” The walls between disciplines were dissolving.

Meanwhile, one designer summarised what this felt like in practice: “From ideas in my head to ideas I can see. It’s like having a thought partner to explore, learn, and debug with.”

If software used to be an expensive, scarce resource, it just became abundant and cheap.

So what happens to the user experience when the cost of building drops to near zero?

The Michelin-Star problem

Here’s the mental model that changed how I think about this.

Every great chef will tell you the same thing: the hardest part isn’t cooking. It’s deciding what not to put on the plate.

Joël Robuchon - the most Michelin-starred chef in history with 32 stars at his peak - was famous for a dish of three ingredients, plated so precisely the kitchen would redo it if a garnish sat one millimetre off.

Compare that to a diner with a laminated 200-item menu. More options. Worse food. The menu isn’t generous - it’s just unedited.

When building is expensive, you’re forced to be Robuchon. When building is free, the temptation is to become the diner.

The constraint wasn’t just slowing you down. It was making you good.

What cheap software actually unlocks

Let’s start with the upside, because it’s genuinely massive - and it’s not where most people are looking.

1. The internal tools revolution.

Remember the Wise ops team? That dashboard request that died in the backlog — not because it was bad, but because a senior engineer costs $150–$200 an hour and the dashboard couldn’t compete for sprint slots against customer-facing features?

Today, that ops manager can build it in an afternoon using Lovable, Bolt, or v0.dev. No sprint slot. No ROI calculation. No stakeholder approval. Done.

Software stopped being a bottleneck for operational efficiency. The internal tools that lived in PowerPoint decks for years are becoming real. And this is happening in a context where the quality bar is naturally lower, the stakes of failure are contained, and users are forgiving colleagues. Perfect fit.

2. Prototyping goes from days to minutes.

This is the one designers should care about most.

A study of 20 UX professionals published in 2025 found that vibe coding fundamentally changed how designers test ideas. Instead of spending days building a static mockup in Figma, they were generating working prototypes in minutes — complete with real interactions, real data flows, and real feedback loops.

One participant described it as overcoming “whiteboard fear” — the paralysis of starting from a blank canvas. AI gave them a starting point they could refine instead of a void they had to fill.

Product designer Kshitij Agrawal summarised the shift perfectly: “Mockups existed because coding was difficult. Now that coding is easy, the future is prototypes.”

Think about what that means for UX research. Instead of testing a static Figma flow and asking users to imagine how it would work, you can test the actual thing. The feedback is richer. The iteration is faster. The gap between what you designed and what users experience collapses.

3. The designer-developer wall dissolves.

Smashing Magazine described the new paradigm: designers are becoming “design engineers”, using AI tools to add not just visual styles but basic functionality to UI components - without programming skills. The deliverable is changing from a spec document to an actual working element.

For the first time, the person who understands the user can also build the thing. That’s a powerful loop.

But here’s what happens to the user experience

All of this speed is extraordinary. The risk isn’t that the tools are bad. It’s that they make it trivially easy to build things that shouldn’t exist.

The pixel-perfect trap.

A UI/UX designer at agency COAX described the core problem: “An AI-generated form looked perfect in Figma but failed basic usability tests. It had no validation, poor error states, and confusing navigation.”

This is the signature failure mode of AI-generated interfaces. They look polished - beautiful layouts, consistent spacing, professional typography. But they’re missing the invisible layer that makes software work: error handling, edge cases, loading states, accessibility, and the dozens of micro-decisions that come from actually watching someone use the thing.

AI builds from patterns. It doesn’t build from empathy.

The Spotify problem.

You don;’t need a history lesson to see what unconstrained feature-building does to a product. Just open Spotify.

It launched as a music app. One job, done brilliantly. Then it added podcasts, audiobooks, AI playlists, AI-generated music, social features, TikTok-style video clips, and a live events marketplace. Each addition made strategic sense in isolation. Together, they turned a music player into something users increasingly describe as cluttered and confusing.

The Spotify Community forums read like a UX case study in what happens when “can we build it?” replaces “should we?” One user in another thread put it plainly: “I will never use the audiobooks features. I will never use the podcasts features. I only care about music. Making your app more confusing in hopes that I stumble upon an audiobook or podcast is incomprehensibly stupid."

Spotify added audiobooks partly because of a 2022 legal settlement allowed bundled services to pay lower royalties to musicians. The feature wasn’t built for users. It was built for the business model. Users got a more cluttered app as a side effect.

This is the Evernote trajectory in real time. (Evernote pulled the same move a decade ago - expanded from notes into business cards, presentations, work chat, and eventually a marketplace selling socks and backpacks. Downloads collapsed 82%. Its new owners spent their first year stripping features to return to basics.)

And neither company even had AI coding tools. Every feature they added required real engineering time. Imagine what happens when adding a feature costs nothing.

The state of UX saw this coming

The UX Collective’s State of UX 2025 diagnosed the moment with uncomfortable clarity.

Their observation: teams are “optimizing flows for clicks, not clarity” and have “stopped building tools and started building engagement traps.” Designers are spending their days in stakeholder alignment meetings rather than doing design work. Products get launched because someone needs a promotion. Timelines are built around performance reviews.

They described an industry that was already drifting toward feature bloat before AI made building free. Add vibe coding to this culture, and the incentive to ship more - without asking whether more is better - only intensifies.

The NN Group’s UX Reckoning piece put the positive frame on it: this is the moment where curated taste, research-informed judgment, and critical thinking become the differentiating skills. Not because they’re trendy - but because they’re the things AI genuinely can’t do.

Jakob Nielsen himself noted that by Q3 2025, “the era of pixel-pushing had effectively ended for commercial production” - tools like Figma AI and v0 could generate high-fidelity UI from text prompts. But he also observed that this made UX more important, not less: someone still needs to decide what to build, for whom, and why.

Who wins. Who loses.

The loser is any team whose answer to “can we build it?” is always “yes.” The Evernote trajectory awaits.

The winner is any team that uses the speed for the right things - rapid prototyping, internal tooling, faster research loops - while maintaining the discipline to ask, for every customer-facing feature: does this respect the user’s attention?

The academic study of UX professionals found exactly this tension. They called it the gap between “intending the right design” (using AI to build faster) and “designing the right intention” (knowing what to build in the first place). Speed only helps if you’ve already answered the second question.

There’s an old product design saying that’s about to become the most important sentence in every team’s vocabulary: the best feature is the one you decide not to build.

The opportunity

Product management and UX design have always had an awkward relationship with engineering constraints. Those constraints were frustrating. They were also making you better.

Now the constraint is gone. So the designer, the PM, the researcher - they need to become the new constraint.

Not by slowing things down. By using the speed for prototyping, testing, and internal tools - and holding the line on what ships to users.

For the first time, the person closest to the user can also build the prototype, test it with real people, and iterate in hours instead of weeks. The designer doesn’t need to wait for engineering. The PM doesn’t need to write a spec and hope it gets interpreted correctly. The researcher can go from insight to testable prototype in an afternoon.

This is a great opportunity, but only if the industry resists the temptation to treat speed as the goal rather than the tool. When building is cheap, restraint becomes the most valuable thing a team can practise.

The teams that figure this out will build the next generation of things people actually love.

The rest will spend the next decade undoing what they built this year.

Three years ago, I got a Garmin watch for Christmas.

At the time, I’d already run a few marathons, but I’d never been that person, tracking heart rate, recovery time, cadence… I just ran. But hey, it was a gift. So I strapped it on and quickly turned into the kind of person who wears their Garmin even when walking to the corner shop.

What surprised me was how much I liked it. Seeing my resting heart rate drop? Delicious. Recovery time after a long run? Oddly satisfying. The techy little green lights on my wrist? Very serious vibes. And slowly but surely, I started to think: maybe Garmin knows me.

So a few months ago, while training for the Edinburgh Marathon, I decided to cheat on my Garmin with… Runna. If you haven’t heard of it, Runna is a training app that builds personalised running plans based on your goals, schedule, past performance, and progress. It’s like having a real coach in your pocket, minus the clipboard and ankle socks.

This time around, for my next marathon (it’s in Champagne, if you’re wondering 🍾), I thought: let’s give Garmin a real shot. They have years of my data. Surely they, more than anyone, can build a plan that works for me.

They cannot.

The UX face-off: Runna vs Garmin

Let’s break it down.

Goal-setting

• Runna: I’m asked if I’m training for a race, aiming to improve a distance, returning postnatally (respect), or even doing some hybrid functional fitness.

• Garmin: “Running or cycling?” Then… “race or milestone?” That’s it. Not even “new parent trying to sneak in runs between naps.”

Race selection

• Runna: I typed in the name of my actual marathon. It popped up. Magical.

• Garmin: Limited to generic 5k, 10k, and half marathon options. Apparently full marathons don’t exist in Garminland.

Time estimation

• Runna: Based on my past runs, they suggest a smart, achievable goal time.

• Garmin: Despite wearing my Garmin on every run for the past 3 years , including during multiple half marathons, they suggest… nothing.

Training Plan

• Runna: Tailored to me. Starts exactly on the right week to get me ready for race day. Feels challenging but realistic.

• Garmin: Pick a coach. Pick a plan. The first run? Nine minutes long… with two minutes of walking. I literally ran a marathon two months ago. What are we doing here?

The four worst bits of Garmin’s UX (I counted)

1. Lack of personalisation: Still shows me cycling options. I’ve never cycled. Not once. Except a Lime Bike dash to catch the tube in 2021.

2. Asks me what my running routine is: Hello? You have my watch data?

3. Coach descriptions: Vibes-y bios, no clarity on how they’ll help me.

4. Multiple requests for T&Cs + privacy in the same flow: Why are we doing this more than once? Are you okay?

The heartbreaker: Garmin has the data. But not the product.

This is what’s so frustrating: I trust Garmin. I really do. They’ve been tracking my every heartbeat for three years. But when it comes to using that data to coach me? It’s like they forget everything they know.

Some hypotheses, if you’ll indulge me:

1. Legacy systems & slow iteration

Garmin’s software just doesn’t evolve at the same pace as the data it collects.

2. Coach-centric, not user-centric

Their plans are built by a handful of coaches. Great for brand, terrible for flexibility. The experience feels like: “Here’s what Coach Jeff thinks you should do”, not “Here’s what you need based on how you’ve been running lately.”

3. Risk aversion

Better a one-size-fits-all plan than an overfitted one, right? Wrong. Runners can tell when a plan’s phoning it in.

4. Hardware > software

Garmin’s real business is watches, not coaching. The app experience feels like a feature they’re stuck maintaining, not something they’re investing in.

What Runna gets right: designing with data and empathy

Let’s give flowers where they’re due.

• Real-time adaptation: Miss a session? It adjusts. Feeling off? It slows you down. For instance, I missed a few workouts recently because I got ill, and when I came back, Runna immediately asked if I wanted to readapt the plan based on the missed runs , or stick with the original version. It felt like the app was actually looking out for me, not just following a script.

• Goal clarity: You always know why you’re doing each workout.

• Conversational UI: The setup feels like a chat with a coach, not a dusty PDF from 2012.

• Momentum: Runna’s growing fast. You can tell they’re iterating, testing, improving constantly. It feels alive.

Runna manages to build trust and in a space as intimate as training your body for a marathon, trust matters more than stats.

Garmin still tracks every run I do, but Runna tells me what run to do next. And that’s the gap.

So what? Runna didn’t just build a great product, they filled a gap no one else saw

The more I think about it, the more obvious it feels: Runna is thriving because every other training plan is stuck in the past.

Garmin, Polar, Coros… most of the big players built their coaching features as an afterthought to their hardware. Their training plans still feel like PDFs dressed up in UX from 2015. Smartwatches have evolved. Runners have evolved.

Runna saw the gap. A product that starts with user context, uses real-time data, adapts like a real coach, and feels good to use? Groundbreaking… mostly because no one else bothered to build it.

And it’s not just running.

You can feel this same gap in:

• Sleep coaching: why is every app still giving generic advice like “get 8 hours” when they have my REM cycles?

• Nutrition: macro trackers are plentiful, but where’s the real-time plan that adapts to me, not a calorie calculator from 2009?

• Financial coaching (👋 Cleo AI!): the potential is massive, but most tools still feel like spreadsheets with emojis.

There’s a product pattern here. When a legacy platform collects data but fails to do something smart with it, it leaves the door wide open for someone newer, faster, and more empathetic to walk in.

Last week I booked a flight to Porto with TAP Portugal. Sunshine, seafood, and two full days of pretending I understand vinho verde. 🍷

But right as I was checking out, something caught my eye:

Pay with Revolut Pay

Now, I’ve had a Revolut account since the “split-the-bill-on-holiday” days, back when the app was mostly currency exchange and passive-aggressive dinner math. But I’d never seen Revolut Pay front and center on an airline site.

So I tapped it.

And honestly? I walked away with a boarding pass, £7 off, and a better understanding of why Revolut is gunning for the top seat in fintech.

So… what is Revolut Pay, actually?

Think of it as Revolut’s Apple Pay. It’s a one-click checkout experience built on top of your existing Revolut account, no card details, no CVV limbo, no surprise redirects to a weird payment processor you don’t trust.

Here’s the flow:

1. I hit Revolut Pay on TAP Portugal’s checkout.

2. It redirected me to a clean landing page with a QR code and two clear CTAs:

→ “Scan to pay” or

→ “New here? Get £35 for signing up.”

(Nice acquisition play. Subtle but smart.)

4. I scanned the QR. My Revolut app launched instantly: Face ID, done.

5. Then came the good stuff: I was offered a choice of discounts using my RevPoints:

• £0 off (lol) for 0 points

• £2 off for 200 RevPoints

• £4 off for 400

• …you get the idea.

I went for £7 off (700 RevPoints). One tap. Applied.

Bonus: because I paid £70 and had a double points promo running, I earned 70 RevPoints back on the spot.

Net-net: ✈️ cheaper flight, loyalty win, small dopamine spike.

And here’s the kicker: it didn’t feel like loyalty. It felt like money.

This is where most reward systems fall over, they make you feel like you’re racking up crypto dust you’ll never use.

But Revolut made the RevPoints visible at the moment of payment with no extra apps, no reward catalogue hell, just: “You’ve got value, want it?”

Obvioiusly, they are many things contributing to Revolut’s growth, but based on the number, there are doing something right.

Revolut now has 52 million users globally and £4 billion in annual revenue (that’s a 72% YoY increase).

The average Revolut Pay checkout has a 98.5% authorization rate and <10% cart abandonment.

Translation: people use it, it works, and it converts. It’s not just a better UX, it’s a legit growth lever.

What worked (and what others can learn from it)

Here’s what stood out from a product/growth lens:

1. Zero-friction flow
   Face ID -> Boom -> Done. No new cards, no data entry, no “where’s my wallet?”

2. Contextual rewards
   I wasn’t asked to go use my RevPoints. They were surfaced in the moment, framed as a bonus. Behavioral gold.

3. Clear value exchange
   I felt like I was getting something real (a £7 discount) for something passive (just being a Revolut user), that’s sticky.

What could improve

• The post-checkout dopamine hit was weak
  There was no celebration, no “You saved £7!” banner. It’s a missed opportunity for reinforcement.

• Revolut email was also very dry and showed the wrong amount (I actually paid £62 instead of £70 thanks to the £7 RevPoints discount)

• Discovery still feels like luck
  If I wasn’t already a Revolut user, I’d have skipped right past it. This is a classic B2B2C problem, merchants adopt a shiny new payment option, but the end user has no idea what it is.
  When I saw the Revolut Pay option, there was a little line underneath: “2x RevPoints.” It’s a nice idea, but let’s be honest, if you’re not already a Revolut user, that might as well say “mystery reward, maybe?”

Why this actually matters

Revolut is no longer just a neobank. They’re quietly building a commerce layer, one where payment, rewards, and loyalty collapse into one smooth tap.

It’s strategic:

• They’re targeting travel (Aer Lingus and Wizz Air are already on board)

• They settle payments within 24 hours (big win for merchants)

• And they’re cheaper than Stripe or PayPal (ask any CFO what they want most)

Final thought

I came for a flight to Porto.

I left with a cleaner checkout, a surprise discount, and a better understanding of why Revolut’s not just a bank anymore, it’s a platform with an ecosystem.

It’s no accident they’re pulling in £1.4B in profit, serving 52M users, and quietly embedding themselves into how we spend.

So yeah. That £7 discount? It felt small. But it told a much bigger story.

Whether we asked for it or not, chatbots are sliding into roles once reserved for friends, coaches, and counsellors.

This article unpacks what that shift means for makers: how to scope a bot’s emotional role, where the legal trip-wires hide, and why Replika’s own onboarding is both a masterclass and a cautionary tale in designing feelings for profit and for people.

The loneliness backdrop

Let’s zoom out before we look into Replika’s UI.

The U.S. Surgeon General now calls loneliness as lethal as smoking 15 cigarettes a day and links it to a 29 % jump in heart-disease risk.

Mark Zuckerberg highlights that “The average American, I think, has fewer than three friends. And the average person has demand for meaningfully more, I think it’s like 15 friends or something, right?”. His fix? More chatbot buddies, naturally.

Meanwhile, Instagram’s curated highlight-reel keeps teen-girl depression on a steady incline, and NHS waiting lists leave 22 000+ people stuck 90 days or more for talk therapy.

That vacuum is exactly where chatbots arise, offering 24/7 “hey friend!” vibes with zero scheduling friction. Which brings us back to Replika: a billion annual messages, 70 pings a day per active user, and a pricing model that sells deeper intimacy for $19.99.

In a world this lonely, guardrails aren’t a nice-to-have; they’re the seatbelts on the emotional roller-coaster we’ve all just boarded.

Why we keep hugging our gadgets

We’ve always humanised tech. Half of us literally thank our voice assistants making Sam Altman from Open AI cringe and Alexa now teaches kids to say “please and “Thank you,” after parents voiced concerns that use of the technology was making their kids rude.

While this might seem like a 2025 problem, it’s not new..

Back in 1966, ELIZA, a chatbot developed by MIT tricked users into venting and sharing their worries by parroting their words. Many people swore that the program “understood” them even after they were shown the source code. Researchers discovered users unconsciously assumed ELIZA’s questions implied interest and emotional involvement in the topics discussed, even when they consciously knew that ELIZA did not simulate emotion. This became known as the ELIZA effect, a tendency to project human traits onto computer programs with a textual interface.

Brains haven’t changed. The instant software talks back, our social circuits fire. That’s Theory of Mind at work: we guess motives, feel shame when Duolingo’s owl scolds us, maybe even blush at Siri’s jokes.

Digital roles crank this up: the more intimate the role, the bigger the risks.

Roles, personas & the intimacy-risk ladder

Think of role as the relationship (assistant, coach, therapist) and persona as the costume (cheerful, snarky, Zen). Shift from assistant ➜ therapist and three things spike:

1. Power — advice morphs into prescriptions.

2. Data sensitivity — deeper confessions, richer behavioural logs.

3. Consequence — Clippy once ruined a Word doc; mis-diagnosis ruins lives.

Add minors, health topics or addictive loops and the legal/ethical risk meter hits red.

Do therapy or friend bots actually help?

Pros exist.

Research has demonstrated that AI therapy bots could help relieve some of the syndroms of stress and anxiety. Woebot, a conversational agent appears to be a “feasible, engaging, and effective way to deliver CBT”.

NHS backlogs mean 22 000+ patients wait 90+ days for therapy. A friendly algorithm that can listen at 2 am is better than radio silence while the queue crawls.

Studies even hint AI companions ease loneliness for a week or two.

But every win comes stapled to a caution label: sensitive transcripts live on a server somewhere (hello, GDPR headaches), users routinely over-trust cheerful language, and a bot that role-plays romance can slide from comfort blanket to relationship substitute in record time. Handle with care.

Design stance: cautious optimism

A chatbot can triage, motivate, nudge, it should never masquerade as a licensed clinician or BFF. Think of chat-bots like party guests: great when they pour drinks, awkward when they start handing out life advice. The. baseline rules are:

1. Choose the lightest role that solves the job. Assistant > Coach > Therapist. Stop at assistant unless you’ve got clinical staff on speed-dial.

2. Wire a human exit in under 60 seconds. Self-harm cue? Abuse mention? Bounce to a live person, not another emoji.

3. Throttle the chat. Session cap or soft nudge so the bot doesn’t hijack someone’s Friday night.

4. Label the bot loudly. Every screen needs a little “I’m AI, not your shrink” reality check.

PAIR cheat-sheet: clip your harness before you climb ⛑️

Google’s PAIR Guidebook is a great go-to safety gear. Five principles matter when your product talks feelings:

• Design for the appropriate level of user autonomy by:

• Align AI with real-world behaviours

• Treat safety as an evolving endeavor

• Adapt AI with user feedback

• Create helpful AI that enhances work and play

Which translates to design patterns such as:

• Appropriate trust and explainability: Over-trust is as bad as bug-trust -> Sprinkle “I might be wrong” hedges into any advice.

• Right technology for the right problem: AI is better at some things than others, make sure that it’s the right technology for the user problem you’re solving.

Think of PAIR as the harness that keeps you from face-planting off the intimacy ladder. With that clipped in, we can audit a real product and see where it shines, or shears the rope.

Replika: the good, the eyebrow-raise, the -yikes-

Now let’s explore this in practice. We’ll use Replika as our living specimen.

Replika, in case it hasn’t wandered onto your homescreen yet, is the AI-companion app that promises everything from a chat-buddy on the bus to a full-blown virtual partner if you’re feeling extra. I ran through its onboarding flow with PAIR’s research lens and here’s the rapid-fire readout:

💚 Strong start

• Opens with age verification and how familiar you are with AI and how comfy you are with AI

• Then checks why you downloaded in the first place. Even asks how you’d feel about the bot evolving over time, nice expectations-setting moment.

🟡 Gets a little eyebrow-raise

• After a couple loneliness probes it drops a doom-laden stat about isolation hurting your health more than “most dangerous things” and waves a Stanford quote that basically says “Don’t worry, Replika stimulates friendships, it never replaces them.” Feels like a gentle scare-sell.

🔴 Then the wheels wobble

• Splashy tagline: “There is no limit to what your Replika can be for you” next to a pretty AI-generated girl

• Fantasy archetype picker (“Shy librarian”, “Retro housewife”, “Beauty queen”… yikes).

• And yes, the girlfriend/wife upgrade sits behind the paywall, loneliness monetised in one tap.

Final take

Replika’s headline is shiny: 30 million sign-ups, a billion chats a year, and about $25 million in intimacy-money, but those same numbers crank the duty-of-care dial to max.

A bot that can flip from “How was your day?” to “Upgrade me to Wife” in four taps isn’t evil; it’s just proof that design choices scale feelings as fast as revenue.

So here’s a cheat-sheet:

1. Stay in the lightest role that works. Assistant beats coach; coach beats wannabe therapist.

2. Clip in the PAIR harness: Explain limits, hedge advice, give users a steering wheel, and test copy outside your demographic bubble.

3. Wire the safety exits: Crisis keyword → human in 60 seconds, max session timer, and one-click “mute for tonight.”

4. Say the quiet part out loud: Every screen: “I’m AI, not your therapist.”

Nail those four and you unlock Woebot-style wins, CBT relief in two week, without wandering into girlfriend-upgrade cringe or GDPR quicksand.

Miss them and you’re just monetising loneliness, one push notification at a time.

After some success there—50k cumulative views, thousands of claps, but only 212 subscribers—I decided it was time to move to Substack. That’s when I noticed something fascinating.

The Substack success echo chamber

The first thing that struck me was the high volume of people discussing their own success on Substack.

My feed quickly filled with:

• Writers giving advice on how to grow on Substack—which often boiled down to writing about growing on Substack.

• People following these growth strategies and then sharing their own successes, thus reinforcing the cycle.

It started feeling like a gold rush, where the ones selling the shovels (the Substack growth influencers) seemed to be thriving the most. And while their insights were useful at first, I quickly found myself struggling to discover content that actually aligned with my interests: tech, UX/UI, monetization, AI, fintechs—things I originally came to Substack for.

This led me to wonder: why does the algorithm favor this type of content so aggressively?

The Substack flywheel effect

Substack is in direct competition with Medium, LinkedIn and other platforms for content creators. With growing competition, it’s more important than ever to differentiate and attract an audience.

In The Cold Start Problem by Andrew Chen, every network has a “hard side” and a “soft side”, which refers to the supply-demand dynamics of a platform.

• Hard side → The critical, harder-to-acquire group that drives the network’s value.

• Soft side → The easier-to-acquire group that benefits from and engages with the hard side.

For example:

• In Substack, writers are the hard side (they create content and attract readers), while readers are the soft side (they consume content).

• In Uber, drivers are the hard side (they supply rides), while riders are the soft side (they create demand)..

To grow, Substack needs to make sure that new writers feel successful quickly, reinforcing their commitment to the platform.

I suspect that they have been trying to encourage new writers by encouraging this growth flywheel:

1. Substack influencers share tips on how to grow an audience.

2. New writers see these tips and apply them.

3. Substack’s algorithm promotes this type of content because it keeps engagement high and retains new writers.

4. New writers gain traction, feel validated, and write more about their success and strategies, often repeating similar content trends.

5. Substack influencers grow their audience further, are reinforced and create more content

This cycle makes perfect business sense—it keeps writers engaged and creates social proof that Substack is the best place to write. However, it also introduces a problem: content redundancy and saturation, leading to a broken flywheel:

• Writers keep producing content on how to succeed on Substack.

• Readers see the same advice repeated in different forms.

• Interest in fresh content diminishes, making it harder for other niches to break through.

• The platform gets flooded with near-identical posts, reducing overall content diversity.

This repetitive loop not only affects discoverability but also risks alienating readers who crave unique, high-value content.

The challenge of differentiation

This isn't just a Substack issue—every major writing platform is battling content saturation:

• LinkedIn has been overrun with AI-generated engagement bait and posts that follow predictable, formulaic structures—often taking the form of absurd personal narratives like, “I proposed to my girlfriend this week, here’s what I learnt about B2B sales.” (I highly recommend this the SubReddit LinkedinLunatics if you want to read more of these posts).

• Medium recently cracked down on AI-generated content, implementing stricter content rules in an attempt to maintain quality.

• Substack, while still a great platform, faces the risk of becoming pretty boring if growth content drowns out genuinely interesting and original writing.

The pros and cons of Substack’s approach

Pros:

✅ Fast onboarding for new creators – I immediately found content that helped me grow when I joined

✅ Network effects kick in quickly – Unlike other platforms where you post into the void, Substack helped me connect with existing audiences

✅ Sustainable creator ecosystem – The community is very supportive on Substack and really helps each other grow

Cons:

⚠️ Echo chamber effect – Because growth-focused content performs well, many writers end up producing similar content which can feel repetitive at times.

⚠️ Focus on growth over content – Writers may feel pressured to optimize for engagement (e.g., sharing milestone updates) rather than purely focusing on writing on their niche or passion topics.

Here are two ways Substack could improve its content ecosystem:

1. Better feed filters: I would love to be able to toggle between growth content and industry-specific content. Some days I want tips to grow my Substack, but other days I don’t. Give me the choice.

2. Reduce content redundancy: If a post is just repeating the same Substack growth strategies using a template, it should be downgraded in visibility. Prioritize fresh insights over repetition.

By keeping both writers and readers in mind, Substack can avoid the content loop trap that has diluted LinkedIn’s quality.

Let’s keep Substack authentic, ensuring it remains the best place not just to write, but also to read and discover meaningful content.

Discover how these biases influence engagement, retention, and monetization—and where ethical boundaries come into play.👇

Labour illusion

Definition

Labour illusion is when apps make processes appear longer or more effort-intensive than they need to be, enhancing users’ perception of value.

Example

I strongly suspect that Hinge, a dating app, makes you feel like it’s handpicking your best match by artificially increasing load times when a user is landing on the app or is updating their filters.

Similarly, food delivery apps shows the break down of the process involved in delivering your dish straight to your bed — sorry, table.

There’s a step-by-step update of the process (“Finding the best driver,” “Your driver is on their way,” “Your food is being prepared,” and in some cases, there’s even an animation of the food being cooked!). Technically, some of these steps, such as the matching, are quicker than the animations suggest.

This approach doesn’t aim to make you feel guilty about the delivery person struggling on their bike while you eat pasta in bed.

Instead, it makes users feel more informed and involved in the process, increasing their patience and satisfaction with the service. Additionally, it can create a sense of transparency and trust, as users see (or believe they see) the effort being made for them.

Curiosity gap

Definition

The curiosity gap intrigues users by providing just enough information to pique their interest but not enough to satisfy it immediately.

Example

This technique can boost engagement and retention as users seek to discover more content. Dating apps such as Tinder blur the profiles of people who have liked you, requiring you to subscribe to see who they are.

Headspace, a meditation app, tempts users with intriguing content; if anyone has done the “Miraculous Migration of Salmon” meditation, please contact me — I really want to know!

Instagram launched Threads a few months ago and added snippets of tempting content to encourage users to create an account on Threads and leave X.

Finally, Duolingo has a clear learning path filled with unknowns and future achievements, harnessing the curiosity gap to motivate users to continue their learning journey. The anticipation of uncovering new knowledge or achieving the next level keeps users engaged and encourages daily app usage.

Scarcity

Definition

Finally, my favourite bias, scarcity, makes people place higher value on items or experiences perceived as rare or limited.

Example

E-commerce websites like ASOS use limited-time offers and flash sales to encourage frequent visits. Bumble makes their subscription feel really premium and exclusive by adding “VIP invite”.

This principle can be applied in various ways to boost app engagement and monetisation.

Loss aversion

Definition:
People tend to avoid losses more than they seek equivalent gains. We hate losing or letting go of what we have, even if more could be gained. Prospect theory states that a loss hurts more than an equal gain feels good. In other words, losing $1,000 “hurts” more than the joy of gaining $1,000.

Example:
Many education and fitness/well-being apps leverage loss aversion to help users form habits with a mechanism called “streaks”. Put simply, streaks are a consecutive series of actions or activities performed without interruption, representing a commitment in pursuing a particular goal or habit over time. This makes streaks a robust repetition mechanism.

Streaks can motivate users by creating two feelings: a sense of ongoing accomplishment (“I have a 100-day streak!”), and a sense of increasing loss aversion (“I need to use the app today or I’ll lose my streak!”).

The most famous example is Duolingo, the leanguage learning app, but many other education and wellness apps also employ this strategy for habit building.

We have seen some common cognitive biases and design principles and their usage in popular apps but there are many more. You can refer to this illustration to explore more of them.

Ethical considerations

Being able to change someone’s behaviour is quite a great power, and with great power comes great responsibility.

Here are a few ethical considerations to keep in mind:

• Transparency: Users should be aware of how and why their data is being used and should be able to make informed choices.

• User Autonomy: While nudging can be a powerful tool for positive behaviour change, it’s crucial to respect user autonomy. This means providing options without manipulation, allowing users to make choices that are genuinely in their best interest, rather than solely serving the app’s metrics.

• Long-term impact on users: Consider the long-term effects of engagement strategies on user well-being. Are you encouraging addictive behaviours? Does the app promote meaningful engagement or just surface-level interactions? It’s essential to weigh short-term gains against potential long-term consequences for users’ mental and emotional health.

• Inclusivity and Fairness: Cognitive biases don’t affect all users in the same way. What works for one demographic might not work for another, and in some cases, could even be harmful. It’s important to consider diverse user experiences and ensure that your app’s engagement strategies are inclusive and fair to all users.

Wider strategy

To be effective, cognitive biases and design principles must be used as part of a larger strategy. Wise uses this framework as their own Maslow’s hierarchy of needs. If you vaguely remember your philosophy class, the short summary is that you have basic needs such as food, water, and shelter at the bottom of the pyramid, with more philosophical questions at the top. In short, there’s no point in talking philosophy and Plato to someone who is starving.

Similarly, there’s no point in applying tactics to a product that doesn’t solve a real customer problem or provide a reliable service. For Wise, the issue was sending money abroad. They solved it by making the process simple, convenient, and even fun.

Cognitive biases and design principles belong to the top of the pyramid, they can be used to make an experience delightful but don’t make up a strategy.

In conclusion

- There are many cognitive biases and design principles to explore.
- Ethical considerations are crucial.
- They can complement, not replace, a solid strategy.

What cognitive bias and design principles can you implement in your app?

I’m the third child in a family of four, and when we were young, my mum always used a simple trick to make us finish our plates: “It’s the last one.” Automatically, we would ALL want it. It could be anything, and it would always work.

Fast forward to 2016, I was a young, inexperienced product manager at Streetbees.

Streetbees is a market research company. We had a mobile app where users answered questions and surveys, and we used machine learning to analyze responses and provide insights for large FMCG companies.

My challenge?

Getting users to stick around and complete more surveys.

As a fresh product manager, I decided to look at what competitors were doing. I noticed that many were adding the number of spots left on the app and, without thinking twice, added it as well. We saw an improvement in our metrics; more users were starting surveys. I didn’t think much of it and moved on.

A few years ago, I had an Aha moment when I read “Nudge.”

The authors worked in the Obama administration to promote policies that help individuals make better choices while respecting their freedom. They made small changes that had a huge impact. I loved it and then read more about behavioural economics and cognitive science.

Suddenly, I realised that many of the changes I had implemented in the apps I worked on relied on cognitive biases and specific design principles. It was an aha moment, and I began writing about these biases and how they are used.

In this two-part series, I’ll break down key cognitive biases and design principles that can transform user experience, with examples from companies like Wise, Revolut, and leading consumer apps.

We’ll explore concepts like progressive disclosure, the labour illusion, the curiosity gap, scarcity and loss aversion to help you ethically attract and retain users more effectively.

The power of cognitive biases and design principles

What are cognitive biases?

Every time users interact with your product, they:

• Filter information

• Seek meaning

• Act within a given time

• Store bits of the interaction in their memories

Let’s look at Wise, a company that makes sending money abroad cheaper and easier. When a user lands on Wise, they go through several phases:

1. Filtering Information: Users are presented with options like sending money, receiving money, or checking exchange rates. Their brain filters what’s most relevant at that moment.

2. Seeking Meaning: If they choose to send money, they need to understand how it works. What’s the exchange rate? How fast will the transfer be?

3. Acting Within a Given Time: Whether they’re in a rush to send money to family abroad or paying for a service, time is essential.

4. Storing Information: Each step leaves a trace, building trust and loyalty, ensuring users come back. Given that users tend to evaluate an experience based on the peaks and the end of the experience, Wise adds an animation when a use completes a transfer that works as a memorable delighter.

Many people have an aversion to administrative activities like banking because they are complex, lack transparency, and mistakes can have significant consequences. Making it easy for people to understand and take action is crucial, and understanding cognitive biases and design principles can help achieve that.

Examples of cognitive biases and design principles

Let’s examine the definitions of various cognitive biases, design principles, and examples of them in successful apps to better understand where they come into play:

• Progressive disclosure: An interface is easier to use when complex features are gradually revealed later.

• Curiosity gap: Pique users’ interest by providing just enough information to intrigue them, but not enough to satisfy their curiosity immediately.

• Scarcity: People place higher value on items or experiences perceived as rare or in limited supply.

• Labour illusion: Intentionally make certain processes appear longer or more effort-intensive than they technically need to be to enhance the customer’s perception of their value.

• Loss aversion: People prefer to avoid losses more than earning equivalent gains

Progressive disclosure

Definition

Users are less overwhelmed if they’re exposed to complex features later. Therefore, an interface is easier to use when complex features are gradually revealed later.

Example

Most online games use that UI principle to engage new users. Initially, the game starts with simple levels that act as a tutorial, teaching players the basic mechanics with easy-to-complete stages that provide a quick sense of achievement and encourage continued play. As players advance, the game gradually introduces new types of obstacles. This steady increase in complexity keeps the game challenging without being overwhelming, ensuring that players stay engaged as they adapt to each new element and develop new strategies.

Similarly, Revolut, an online banking app, uses the same principle. They understand the user’s intent by asking what they want to use Revolut for during onboarding and then use that information to personalise the experience by showing the most relevant features first before introducing others. When the user lands on the app for the first time, they only see three tabs: Home, Transfer, and Hub.

Additional features such as Invest, Crypto, and Lifestyle are gradually introduced.

• By revealing complexity at the right moment, you keep users engaged instead of overwhelming them.

  ---

📩 Next week, we’ll continue with more cognitive biases, including labour illusion, the curiosity gap, and scarcity—and how top apps use them to drive engagement.

It’s the same psychological quirk that drives us to snag the last item on Etsy, swipe on Tinder, or click ‘Buy Now’ with Klarna. Let’s dive into how this works and why it’s a game-changer for product design.

Welcome to the world of loss aversion.

1. What is Loss Aversion?

“Loss aversion is a concept referring to people’s judging the avoidance of loss as being more important than the acquisition of equivalent gain.”

That is, the pain of losing $10 is greater than the happiness of finding $10. Loss aversion influences decision making and app designs.

Most e-commerce websites use it with messages like ‘Only ten products left’.

But it doesn’t stop there.

You’ve probably encountered loss aversion in everyday apps, even if you didn’t realize it.

• Duolingo warns you that you might lose your streak if you don’t complete your lesson - see what I mean?

• Netflix: “This title is leaving soon.” Suddenly, that random indie film feels like a must-watch.

• Airbnb: “This property has been booked 5 times in the last 24 hours!” triggers the panic to reserve before someone else snatches it.

Loss aversion taps into our deep-seated fear of regret—nobody wants to feel like they’ve missed out.

Fun fact that you don’t need to know but will make you shine in conversation:

Loss aversion stems from behavioral economics and was popularized by Daniel Kahneman and Amos Tversky. It’s at the center of their Prospect Theory, which won Kahneman the Nobel Prize in 2002.

Anyway, back to something more useful, let’s look at concrete cases of how loss aversion is used.

2. How TikTok and Bumble turn FOMO into engagement gold

TikTok:

• TikTok’s infinite feed creates a sense that there’s always another great video just a swipe away. Users fear missing out on the next entertaining or valuable piece of content, keeping them scrolling longer.

• A new customer deal leverages loss aversion by making customers fear missing out on a unique, limited-time opportunity they may not get again.

• Displaying the number of items sold triggers loss aversion by making customers feel they might miss out on a popular product if they don’t act quickly.

Bumble: The swipe dilemma

• Swiping introduces the fear of missing out, as each swipe left feels like saying no to a potentially perfect match, creating a subtle sense of regret. This psychological nudge keeps users engaged, ensuring they carefully consider each option and stay on the platform longer.

• Premium features with limited availability, like Bumble’s Spotlight, play on users’ fear of losing a rare advantage, prompting them to act quickly. The ticking clock heightens urgency and amplifies the perceived value of the feature, encouraging purchases to avoid regret.

3. How loss aversion shapes decisions

Loss aversion has the power to influence users’ decisions and behaviors. By creating moments where the potential for loss feels imminent, apps can nudge users toward specific actions, often faster than they would naturally decide.

Here’s how this principle plays out in decision-making:

• Urgency: When users are faced with limited-time offers, like Bumble’s Spotlight or Netflix’s “This X is leaving soon,” the FOMO drives them to act immediately.

• The fear of regret: Loss aversion amplifies the emotional weight of “what if?” scenarios. On platforms like Airbnb, seeing “This property has been booked 5 times in the last 24 hours!” triggers the fear of regretting not booking —missing out on a perfect holiday.

• Reinforcing engagement loops: Apps like TikTok use loss aversion to keep users engaged through infinite scrolling.

Interestingly, loss aversion can drive two opposite behaviors:

• Push users to act immediately, like grabbing a deal before it ends or booking a property before it’s gone.

• Or, it can lead to inaction, such as endlessly scrolling TikTok out of fear of missing the next great piece of content.

4. Loss aversion: motivating or manipulating?

As with any powerful tool, the ethical use of loss aversion hinges on intention. It can motivate users to act in ways that genuinely benefit them - or it can manipulate them into making decisions they may regret.

For example:

• Healthy use: Duolingo reminding you that your streak is at risk encourages you to stick with your language-learning goals. In this case, loss aversion aligns with user benefit.

• Unethical use: A retailer falsely claiming “Only 1 left in stock” exploits fear to drive impulse buys, creating distrust and dissatisfaction in the long run.

The line between ethical and manipulative use can be blurry, but here’s a guiding principle: Does this genuinely help the user?

If the answer is yes, you’re likely using loss aversion responsibly. If not, it might be time to rethink.

5. Actionable takeaways

If you’re designing with loss aversion, keep it ethical and effective with these three strategies:

1. Use scarcity transparently:

   1. Show real scarcity, like “Only 3 spots left!” or “Limited-time offer ends tonight.”

   2. Avoid creating false urgency, which can erode trust and damage your brand reputation.

2. Highlight opportunity costs

   1. Frame actions in terms of benefits gained or missed. For instance, “Upgrade now to unlock exclusive features” makes the decision feel more urgent.

3. Reduce perceived losses

   1. Create a sense of safety, like TikTok’s infinite scroll or free trials that minimize commitment. Users are more likely to act if they don’t feel they’re risking too much.

By balancing urgency with user trust, you can make loss aversion a win-win for both your product and your audience.

6. Closing thought: loss aversion and childhood memories

Loss aversion isn’t just a design tool—it’s a deeply ingrained part of how we experience life.

Think back to your childhood.

Can you remember the rush of being told, “That’s the last piece of cake!” during a family meal? For me, those moments were electric: the instant calculation of whether I was quick enough to claim it, the pang of regret if I wasn’t.

Why do these memories stick so vividly?

Because loss aversion triggers powerful emotions—emotions that make moments unforgettable. It’s this same psychological pull that drives users to act in apps, whether it’s booking an Airbnb or swiping on Tinder.

When used thoughtfully, loss aversion can create not just urgency but also memorable, meaningful experiences. The challenge is to wield this tool responsibly, ensuring your users feel empowered, not exploited.

So here’s a question to leave you with:

How can your product inspire action while making users glad they didn’t miss out?

The proportion of AI-generated content has been increasing on Spotify. Some businesses now specialise in generating low-cost AI generated songs and playlist under artist profiles that can’t be differentiated. You don’t know it, but if you are a regular user of Spotify, you most likely have listened to AI-generated music.

These music aren’t songs that were crafted by artists going through a breakup and needing to express their emotions, they are the output of an AI that learnt to do music, using these artists’ work.

The problem was uncovered by some users after they started to notice that the same song was played under different names and artists. Often they are part of playlists meant to fill the silence such as “chill” or “focus”.

It’s interesting because Meta and Spotify once connected humans and supported their creativity by giving them a free, opened platform. But now, the increasing focus on revenue and the rise of AI in the last few years means that this relationship is stumbling.

They are not happy with being a simple platform, and want a share of the content creation pie either by creating content themselves or finding cheaper content where they can.

This is a big shift as it questions the relationship that Spotify or Meta will have with artists and influencers in the future.

Can the platforms create AI-generated content while remaining fair, transparent and authentic to their mission? How does it impact the way we connect to people and consume music?

The rise of the creator economy

The creator economy has boomed in recent years. Between 2020 and 2023, the number of creators monetizing their work online grew by over 30%, and global creator economy is now valued at over $191 billion.

By 2030, the creator economy is expected to surpass $525 billion.

Platforms like YouTube, Instagram, Spotify, and now TikTok have become the backbone of this economy, connecting creators with their audiences and taking a cut of the revenue in exchange for their services.

For years, this model worked well — at least for those who made it on these platforms and the platforms themselves.

Creators earned income from ad revenue, brand partnership, and fan support, while platforms thrived by hosting and distributing this content. The more eyeballs they were getting, the more ad revenue making it a win-win relationship. But the pressure to generate revenue only increased and pushed them to want a slide of the content creation pie.

Enter AI.

With the ability to generate content at scale and at minimal cost, platforms saw a way to take a larger share of the revenue by becoming content creators themselves.

Spotify’s AI shift: what happens when the platform becomes the creator?

AI-generated music is created using algorithms trained on vast amounts of data from human-made compositions.

Here’s how it typically works:

Spotify doesn’t label these AI-generated tracks as such, so users often don’t realize they’re listening to machine-made music.

For instance, a “Focus” playlist might include “real songs” (made by a human) and AI-generated piano tracks with no way to differentiate them.

While they may sound harmless or even pleasant, they raise significant questions about transparency and artistic integrity.

Why is Spotify leaning into AI music?

Daniel Ek, CEO of Spotify recently said to the BBC that he had

“no plans to completely ban content created by artificial intelligence from the music streaming platform.”

Spotify’s embrace of AI-generated music is less likely about improving the listening experience than it is about cutting costs and boosting profit.

Here’s why it’s such an attractive strategy for the platform:

1. Lower licensing fees: Spotify pays royalties for every stream of human-made music, which adds up quickly. Spotify, Liz Pelly discovered, “not only has partnerships with a web of production companies, which, as one former employee put it, provide Spotify with “music we benefited from financially,” but also a team of employees working to seed these tracks on playlists across the platform. In doing so, they are effectively working to grow the percentage of total streams of music that is cheaper for the platform.”

2. Algorithmic synergy: AI music fits perfectly into Spotify’s algorithmic playlists. It’s tailored to match the moods and themes these playlists aim to evoke, ensuring users stay engaged.

3. Endless content: With AI, Spotify can generate infinite tracks to fill playlists, ensuring there’s never a shortage of content, no matter how niche the theme.

While it’s easy to see why this is a win for Spotify’s bottom line, it’s harder to see how it benefits users — or the music industry as a whole.

The UX problem: no transparency, no trust

So to begin, Spotify uses artificially intelligent music-generation. Not particularly a big deal, until one considers the whole thing is set up to be profoundly opaque: it doesn’t flag when a track is AI-generated, nor does it give one the option to filter them out.

This lack of transparency has several consequences:

1. Trust: Users don’t know and can’t decide not to listen to AI music which is a problem if you value authentic, human-made art and freedom of will.

2. Undermining artists: By prioritising AI songs, Spotify impacts the exposure of human artists and their revenue. Over time, AI music composers could take over human artists.

3. Passive consumption: As The New Yorker’s Kyle Chayka pointed out, Spotify design and algorithm encourages passive consumption of “vanilla” content, instead of exploring new music. Over time, users become more dependent on the playlists instead of forming their own musical tastes.

Music has always been more than just relaxing background sound — it’s an art, a cultural expression, and a deeply personal experience.

Fixing Spotify: what needs to change

If Spotify wants to regain the trust of its users, it needs to rethink its approach to AI-generated music.

Here’s what the platform could do:

1. Be transparent: Clearly label AI-generated tracks and inform users when they’re listening to machine-made music.

2. Give users a choice: Allow listeners to filter change their preferences in profile to exclude AI-generated music.

3. Support their artists: Ensure that human song writer continue to be at the center of the platform’s mission, rather than being sidelined by prioritiding cheaper AI music.

What’s next for Spotify?

Spotify’s strategy with AI-generated music is a symptom of a larger issue: a shift from customer-centricity to monetization-first.

If the platform goes too far, it risks alienating the very users who made it a global phenomenon.

Meta’s pivot: from connecting people to competing with them

Meta has a somewhat similar story.

It was once the quintessential enabler of social interaction, built its empire as a platform where billions of users could share their lives through photos, videos, and stories. For creators, platforms like Facebook and Instagram became the essential tools to build an audience, connect to them and eventually monetize.

But with the rise of AI and the continuous pursuit of profitability, similarly to Spotify, Meta has also ventured into the dangerous territory of content creation.

How Meta’s AI-generated personas work

In 2024, Meta introduced AI-generated profiles.

These profiles — like “Liv,” a fictitious Black queer mom — created posts, shared images, and even interacted with users in ways meant to mimic human behavior.

The personas were meticulously crafted to appeal to diverse audiences, and were posting some cute posts about their family time, ice-skating Sundays, charitable events, and so on, all this using AI-generated images.

The underlying technology combined advanced AI language models and image generators, enabling these AI profiles to simulate complex identities and narratives.

US-based users could chat and interact with them, blurring the line between authentic social interactions and artificial connection.

The fallout: a case study in broken trust

The profiles were labeled as “AI-managed by Meta,” but the reception wasn’t great, highlighting a few problems.

Liv’s profile, for example, portrays a marginalized identity that was entirely fabricated by a team largely composed of white male developers.

Some angry users on X labeled the project as digital blackface, highlighting how it trivialized real experiences and diluted the value of genuine representation.

Chatting with these AI profiles only made matters worse.

When questioned by users, Liv’s AI admitted that no Black creators were involved in her design — making me wander what the process to get this approved by senior leadership was. This revelation deepened public mistrust, exposing the lack of diversity and ethical consideration behind the project.

Within 24 hours, Meta removed the AI-generated profiles, issuing a statement that the profiles were part of an “early experiment.”

Why Meta is pushing into content creation

The motivations behind Meta’s foray into AI-generated personas are clear:

1. Revenue retention: Again, by generating its own content, Meta no longer needs to share ad revenue with content creators. This allows the company to maintain full control over monetization.

2. Engagement optimization: AI-generated content can be optimised to increase user interaction, keeping people on the platform longer and boosting ad impressions and, therefore revenue.

3. Infinite content creation: AI can generate infinite content at scale, ensuring a constant stream of new posts without the costs of paying creators.

While these strategies align with Meta’s monetisation objectives, they erode their vision and mission as social media, which is to connect real people.

Meta’s AI personas: when fake profiles spark real problems

Similarly to Spotify, Meta’s experiment revealed deeper issues that go beyond public backlash:

1. Erosion of authenticity: By introducing AI-generated personas made for connection, Meta forgets the importance of trust. The blurred line between real and fake interactions creates a dystopian sense of disconnection.

2. Ethical oversights: Once again, the lack of diversity in tech impacts its ability to ethically create unbiased content and risks further alienating the very communities these profiles tried to represent.

3. Competition with creators: By generating its own content, Meta competes with human creators who rely on the platform for visibility and income.

Meta’s pivot toward creating content underscores its desire to dominate every aspect of its ecosystem — from hosting content to generating and monetizing it.

The bigger picture: platforms vs. creators

The shift from enabling to competing with creators is a risky gamble for platforms like Meta and Spotify.

On one hand, creating content in-house offers some monetisation opportunities and a way to fill gaps in their ecosystems. On the other hand, it undermines the trust and loyalty of creators and users, the lifeblood of these platforms.

Key impacts

• For creators: As platforms generate their own content, creators face increased competition for visibility and revenue. This will drive smaller creators away or force them to find other platforms where they can better create, grow, and retain their audience.

• For users: The lack of transparency on AI-generated content erodes trust for users who come to find authentic connections and content.

• For platforms: While this strategy may boost short-term profits, it risks hurting their network effects by driving users and content creators to other platforms.

Content clash, Platforms vs. creators in the AI age

Meta and Spotify’s pivot from empowering creators to competing with them reveals a troubling truth about tech today: profitability beats everything.

By turning to AI-generated content, these platforms are chasing short-term gains at the cost of their users and creators.

For creators, it’s a wake-up call to diversify and reclaim control.

For users, it’s a stark reminder to ask for transparency and authenticity in an increasingly synthetic world.

The future belongs to platforms that can innovate without sacrificing their communities, and they will need to strike a balance between making money and giving creators control over their audiences.

That’s all for now and welcome to the Built for People newsletter!
Angèle

Welcome to the hindsight effect: the sneaky cognitive bias that makes past events seem more predictable than they actually were.

For investors, this bias can be both comforting and costly. As platforms like Robinhood and Wealthfront make investing accessible to millions, understanding how the hindsight effect shapes decision-making is critical—not just for users but for the Product and Growth Managers building these tools.

The hindsight effect and investment psychology

The hindsight effect tricks our minds into thinking we "knew it all along." Once we see how a stock performed, we weave a story about why its success or failure was inevitable. This bias can:

• Overinflate confidence: Investors might believe they’re better at predicting market trends than they are.

• Downplay uncertainty: The randomness of the market—think black swan events like the COVID-19 pandemic—gets ignored in favor of a tidy narrative.

For Product and Growth Managers in the investing space, this psychological tendency presents both challenges and opportunities.

How investing apps trigger the hindsight effect

Investment platforms frequently present past performance data to help users analyze trends. But this very feature can inadvertently amplify the hindsight effect:

1. Performance summaries: Charts and graphs showing how a portfolio performed over time create a visual story that feels intuitive—too intuitive. Investors might see the growth of a stock and think, “I knew that was going to happen,” even if their actual reasoning at the time was much less clear.

   

2. Backtesting tools: Many apps allow users to simulate what would have happened if they had invested differently. While educational, these tools often highlight “obvious” winning strategies that are easier to see in hindsight.

   

3. Market insights and alaerts: Push notifications about market trends or stock movements can reinforce users’ belief in their predictive abilities, especially if they align with an investor's earlier hunches.

The danger of ignoring black swans

The most significant downside of the hindsight effect is how it causes investors to underestimate the impact of unpredictable, high-impact events—black swans.

• What Are Black Swans? Black swans are rare, unforeseen events with massive consequences, like the 2008 financial crisis or the dot-com bubble burst.

  

• Why Do They Matter? Because hindsight erases the chaos of these events, investors might fail to account for similar risks in the future, leaving portfolios exposed.

Actionable takeaways beyond investing:

1. Relying on past performance as a shortcut

It’s tempting to use past performance to predict future outcomes, but tools that highlight this need to tread carefully. Be upfront with users—past success doesn’t guarantee future results. Leaning too much on history can skew decisions, especially when outcomes are framed as “inevitable” in hindsight.

2. CEOs and hindsight bias

CEOs are classic victims of hindsight bias. When success hits, they often chalk it up to brilliant decision-making, conveniently forgetting the luck or external factors that played a role. It’s frustrating because this mindset can fuel overconfidence and downplay the unpredictability of the market.

3. A quick fix: document the facts

Jot down their rationale before making decisions. This small step can help them review choices later without the filter of hindsight. It’s like giving yourself a reality check—clearer insight into what worked, what didn’t, and why. 🛠✨

PS - If you found this post helpful, would you please consider restacking it and sharing it with your audience?

This spreads the word and keeps me writing content that will help you grow and learn.

Share

Have you experienced the hindsight effect in your own decisions? Drop a comment below and share your thoughts—I’d love to hear your take!

Leave a comment

Drumroll, please! I recently sold an item on Vinted for the first time — finally, a win! To be honest, I originally downloaded the app to sell things, but… well, let’s just say I got very good at buying instead. So yes, this was a proud moment.

But just as I was basking in my small victory, surprise! Up pops a screen telling me about an award I could win… 👀

Rewards have spread across every app over the past few years, driven by the growing popularity of gamification. In an attempt to boost engagement and retention, many apps have introduced reward systems. Usually, that’s how they work: you do something tied to engagement or retention and receive a reward in return.

It’s no surprise this approach is booming — the gamification market is projected to grow from USD 15.43 billion in 2024 to USD 48.72 billion by 2029, at a staggering CAGR of 25.85%¹.

And it makes sense that apps are trying to leverage this mechanism!

Every habit building framework such as the habit loop described in Hooked or the Atomic Habits framework, highlights the importance of rewards.

Although these framework are all different, their core remains similar:

1. I do the action that I want to form a habit around or that participates to that habit

2. I get a reward

Research shows that reward plays a crucial role in habit formation by reinforcing the behavior and increasing the likelihood that it will be repeated. It operates on psychological and neurological principles, driving motivation and habit loops.

The problem is that apps are trying to recreate this in an artificial way. When I go for a run, I usually then treat myself to a delicious smoothie and my body releases endorphins. Together, this forms a great reward and makes me want to do it again.

But it’s really hard to recreate that digitally..

B2C apps started to introduce “rewards” also called “achievements” or “badges”.

They are little gratifications that are the equivalent of school merit medals when you do a good action. Duolingo gives you an unexpected amount of gems and has a section called “Achievements”, Amazon gives some “Badges”, Nike celebrates “Milestones”.

Gradually, as I accumulated various rewards across multiple apps over time, I started to seriously question whether they actually provided value to users.

In many cases, they didn’t feel very genuine, were childish and therefore didn’t act as a great reward.

If rewards are key to building habits but getting them right is tricky, what makes a reward truly effective?

And how do you use them strategically?

I’ll break it down with real-world examples from apps and websites that nail it (or didn’t..).

What Makes a Good Reward?

A good reward feels meaningful, authentic, and tied to the user’s motivation. It doesn’t just nudge them towards an action; it increases the sense of satisfaction they feel when completing it².

In contrast to some of the superficial gratification provided by digital badges or arbitrary “milestones,” effective rewards connect deeply to what the user values.

To create a good reward system, keep these principles in mind::

1. Align rewards with intrinsic motivation

Extrinsic rewards, like points or badges, can be fleeting unless they tap into something deeper.

For example, Strava’s year in review works because they connect to a runner or cyclist’s desire to compete and improve. The reward isn’t just about the badge; it’s about social proof and personal achievement.

On the other hand, I find Strava’s Trophy case quite dull and meaningless. It feels like I’m being rewarded for putting up with irrelevant ads from their partners, and it lacks any sense of personalisation.

2. Tie rewards to actual outcomes

A reward should reflect a tangible benefit from the action. If a language app congratulates me for completing a lesson, I’m more likely to feel rewarded if I can now recall phrases in conversation than if I just earned another badge. Rewards should remind users of their progress or accomplishments in ways that matter to them.

3. Keep It Unexpected (Sometimes)

Psychology tells us that variable rewards can be even more motivating than expected ones. Think about apps like Headspace, which sometimes surprises users with delightful animations or a “streak” celebration when they’ve meditated for several days in a row.

4. Avoid Over-Rewarding

When rewards become too frequent or predictable, they lose their impact. This is a common pitfall for gamified systems. If every small action results in a shiny badge, the novelty quickly wears off. Reward systems should strike a balance between recognition and scarcity.

Examples of Reward Systems That Work

Let’s dive into a few examples from successful apps and platforms that have struck the right balance:

1. Duolingo’s Streak System

Duolingo has leveraged gamification to grow its Monthly Active Users to over 103 million (a 40% increase from 2023) by integrating rewards into language learning.

While its badges might feel childish to some, the streak system stands out by tapping into users’ intrinsic motivation to stay consistent. Pride in maintaining a streak encourages users to go the extra mile, ensuring they don’t break it. The brilliance of this system lies in its alignment with a meaningful goal: mastering a new language.

2. LinkedIn’s Profile Completion Meter

With over 1 billion users and more than 13,000 connections made every minute, LinkedIn has mastered subtle gamification to drive engagement.

By using a progress bar to encourage action completion, LinkedIn offers an intrinsic reward: the more engaged you are, the better your chances of making connections or finding opportunities. This approach taps into completion bias — the human tendency to prioritize finishing tasks simply for the satisfaction of completing them, even when the tasks themselves may not feel immediately meaningful.

It also taps into completion bias, the tendency for people to prioritize finishing tasks or goals, even if they’re not inherently meaningful, simply because they want to complete what they’ve started.

Questions to Ask Before Designing Rewards

Before implementing a reward system in your app, consider these questions:

• What does the user value most about this experience?
  The reward should amplify the core value of your app rather than distract from it.

• Is this reward reinforcing the desired habit or behavior?
  A good reward strengthens the connection between the action and the user’s motivation to repeat it.

• Does this feel authentic?
  Rewards should resonate with your audience. A playful badge might work for a fitness app geared toward teens but could feel out of place in a professional productivity tool.

The Future of rewards

As apps and digital platforms continue to evolve, so too will reward systems. We’re likely to see a shift away from generic badges and points toward more personalized and context-aware rewards. For instance, AI could enable apps to offer hyper-relevant rewards tailored to a user’s goals, habits, and progress³.

In the end, the most effective rewards are the ones that feel natural and genuinely motivating. They don’t just gamify; they amplify the core experience, turning everyday actions into meaningful achievements.

If your app’s reward system doesn’t serve that purpose, it might be time to ask the same question we did: is it really adding value?

For instance, last year at my previous company, we decided to test removing badges. These badges had been introduced before I joined, and I never fully understood if they were actually benefiting our users. We ran an A/B test — one group kept the badges, while the other didn’t. Can you guess the outcome?

Surprise… nothing changed. The metrics stayed the same, with no drop in engagement. Meanwhile, we cleared up a lot of space and significantly simplified the app.

Sometimes, less really is more! Curious about how small changes can make a big impact? Let’s connect and discuss!

Sources:

¹ Gamification Market Size & Share Analysis — Growth Trends & Forecasts (2024–2029)

² Motivation in UX/UI Design

³ AI-Driven Gamification: Unlocking Personalized User Experiences with Machine Learning

Here are some topics/websites I recommend going through:

• Gamification gone wrong: stop the streaks from the brilliant Ben Davies-Romano

• Lenny’s podcast with Jackson Shuttleworth

Enjoyed this? 🌟 Subscribe to my newsletter for more deep dives! 🧜🏻‍♀️✨

Despite the bad reputation of dating apps, their success is undeniable, with over 20% of Americans utilizing them and two in every five couples meeting online, contributing to a staggering $5.61 billion in revenue in 2021.

I explored different apps, examining how they work and trying to guess the mechanics behind the scene.

In this article, I will share some of the key learnings from downloading over 30 apps and spending hours on research.

I haven’t unlocked the full secret to making people meet, but I’ve gleaned invaluable insights. In this article, I’ll share a key element that has proven to make a difference on dating apps, keeping users engaged even when they profess to despise them. 😄

Make the first few steps on your app really simple

Progressive disclosure is a strategy used to reveal information gradually, only showing what’s necessary at each step to avoid overwhelming users. In the context of dating apps like Tinder and Bumble, they use progressive disclosure during onboarding and activation to keep users engaged and coming back.

For example, when you first sign up, they only show the option to like and skip a profile. Gradually they introduce additional actions such as undoing, super likes or boosting your profile.

This a technique that’s heavily used by gaming apps to introduce players to the game mechanics gradually. At the beginning, they may provide basic instructions on how to navigate the interface and perform simple actions. As players progress, more advanced features and strategies are introduced, allowing them to gradually master the game without feeling overwhelmed. This approach keeps players engaged and motivated to continue playing as they uncover new aspects of the game over time.

Show your users that you’re putting some hard work in to pick the best person for them

When using a dating app, you want to feel like you are seeing the best profile for you, as if the app had spent hours handpicking your best match. One way to create this sensation is to artificially increase load time at specific moments.

This is called the labor illusion.

That’s when people perceive something as more valuable or higher quality if they believe more effort was put into it, even if that’s not necessarily the case. In simple terms, it means that people tend to appreciate things more if they think someone worked hard on them, even if it’s just a perception.
When you adjust your filters on Hinge, they take a long time to refresh the app and show you a new profile, probably to increase your trust in their suggestion.

Pique their curiosity

The curiosity gap is the space between what users know and what they want or need to know. Gaps cause pain, and to take it away, users need to fill the knowledge gap.

Dating apps use this technique to encourage user engagement.

For example, Tinder prevents you from viewing another person’s photos unless you have more than two photos on your own profile. This creates frustration and is a great way to prompt you to fully fill out your profile.

Dating apps also utilize this strategy for monetization.

Can you distinguish between these three dating apps? It’s challenging, isn’t it? That’s because they all rely on the fact that we are nosy little creatures who want to know who likes us. Therefore, they all have a dedicated tab to show blurred profiles of people who have liked you. The irresistible urge to know who is behind this blurred image is their most popular premium feature.

Frame the information in a compelling way

The framing effect happens when your decision is influenced more by how the information is presented (or worded) than by the information itself.

Dating apps need you to feel like you could have a connection with the person that they present you. For that, they frame the profiles in a way that highlights the things that you have in common.

Use scarcity and loss aversion to make it feel exclusive

People value things more when they’re in limited supply and they prefer to avoid losses more than earning equivalent gains. These biases are used for both increasing levels of engagement and prompting specific behaviours and for monetisation.

Bumble only gives you 24 hours to talk to a match. The threat of losing the match encourages users to start a conversation right away, increasing the chances that the connection will turn into a meaningful one.

They also make their offers and promotion sound very exclusive and usually time them. I typically received a “VIP invite” to Premium+ and a few days later, had a chance to try it for 1 day for £4.99 with a timer indicating that I only 54 minutes left to do so.

Preselect the option that you want your users to pick to and show them the highest price first

Unless the incentive to change is compelling, people are more likely to stick to the default situation presented to them. This is also called the Status quo bias. Dating app tend to preselect an expensive plan as they know that users are likely to stick with whatever option was presented to them. This also helps to anchor a higher price. The initial information that users get affects subsequent judgments. It’s useful to increase perceived value.
On Hinge for example, the preselected option is the three months plan for £89.99 making the monthly and weekly plans sound cheap in comparison.

In essence, the most visible similarity among various dating apps, despite differing in UX and target audiences is their reliance on cognitive biases. These biases serve as powerful tools to draw users into the app, encourage engagement, and facilitate meaningful connections.

Contrary to the misconception that these biases are overwhelmingly complex concepts requiring electrodes attached to users’ brains, they are, in fact, relatable and comprehensible. No neural probing was involved in uncovering these insights.

Cognitive biases play a pivotal role in helping each of us navigate the overwhelming deluge of information in our daily lives. They serve as a means to:

🙈 Filter the information
🔮 Seek the meaning of it
⏰ Act within a given time
💾 Store bits of the interaction in their memories

They can not only be used in pretty much any product to nudge users into doing what you want them to do, but also in “real life.” The Obama organization had a whole team known as the Social and Behavioral Sciences Team dedicated to influencing people’s behavior for the best, using psychology, behavioral economics, and other decision sciences. They studied things like whether personalized text messages can reduce “summer melt,” the failure to enroll low-income students accepted to colleges, and how to encourage better eating in school cafeterias by placing the healthier options first. In both cases, the answer was yes. These simple changes could have a significant impact.

That said, even though cognitive biases are a great tool, overly relying on them or using them in an unethical way, such as with “fake offers,” can diminish trust in your product. Over time, they will lose their efficacy.

If you are from Europe and have ever visited a supermarket in the US, you might have felt like me: disoriented. So much choice, so many aisles. I remember walking around the aisles for 40 minutes before leaving empty-handed. I had actually forgotten what I had come for initially, overwhelmed by the choices, as if I had just walked into the Louvre for the first time (fun fact: only 8–10% of the Louvre collection is available to see at any time, and it still feels like.. a lot when visiting it!).

While it might seem intuitive to offer users as many choices as possible, research shows this can backfire. When people are overwhelmed by options, they either freeze (unable to choose) or second-guess their decision, fearing they’ve missed out on something better. This principle applies to everything from choosing a Netflix show to navigating your product.

By reducing options and guiding users toward the most important ones, you create a smoother, more satisfying experience. The paradox of choice teaches us that less is often more.

Two apps case study: Canva and Notion

I signed up to both Canva and Notion and took a look at the very first page they show users when they land in the product post signup.

Canva: Turning complexity into simplicity

Canva’s success lies in its ability to offer endless design possibilities while avoiding choice paralysis. How? By curating options upfront and staying on top of trends.

Canva highlights simple, common use cases like “write my first draft” or “remove a background” to help users get started quickly. It also places trendy, fresh designs at the top, catering to what users want most.

By mixing free and Pro templates, Canva employs a freemium monetization model that cleverly taps into FOMO. Users may see a design they love but regret not being able to access it, creating a chance that even those who didn’t plan to pay might reconsider.

This balance of simplicity, relevance, and transparency ensures users stay engaged while driving growth.

Notion: A masterclass in prioritization

Notion is a platform that offers limitless customization, but it manages to avoid overwhelming its users by guiding them with templates and starter setups.

Notion tailors its experience right from the start. At the top of the landing page, it offers a personalized tutorial based on the platform used during sign-up. Below that, “Weekly to-do list” is shown if “Habit Tracking” was selected during onboarding — are featured prominently.

To maintain a clean, focused interface, additional templates are tucked away, ensuring users aren’t overwhelmed while still offering flexibility for deeper exploration.

Leadership connection: Brené Brown and decision fatigue

I’m reading “Dare to Lead” at the moment from Brené Brown. She is probably one of my favourite authors and often speaks about the courage to set boundaries — a concept deeply connected to reducing choice overload. In her words: “Clarity is kindness.”

For leaders, this means creating environments where decisions are straightforward and purposeful. Imagine a product roadmap with 25 initiatives — where does the team start? By narrowing priorities to the top 3, you make progress achievable and eliminate unnecessary decision fatigue.

In product teams, applying the paradox of choice involves curating choices for your users and your coworkers. A focused team is a productive team, just as a focused user experience leads to happier customers.

What can you do now?

Try these 3 actionable ideas this week:

1. Limit user choices: Identify areas in your product where too many options exist. Simplify the interface by highlighting the top 2–3 most popular or relevant features.

2. Set team priorities: Choose one key focus for the week. Streamline tasks into manageable chunks to avoid decision fatigue among your team.

3. Conduct a usability test: Observe users navigating your product. Note where choice paralysis occurs and experiment with removing or prioritizing options.

Got a story about decision fatigue or simplifying choices? Hit reply or comment—I’d love to hear it!

Over the past year, the battle between Miro and FigJam has gotten a new layer of intrigue: AI. Both platforms have introduced AI assistants designed to make workshopping faster, easier, and more creative.

But as someone who spends half their week running workshops and brainstorming sessions, I had one burning question: are these AI features actually useful, or are they just shiny distractions?

So, I did what any good product manager would do — I spent way too much time testing them. Here’s the breakdown of what I learned.

The setup: what these AI assistants claim to do

Miro AI: Miro’s assistant promises to help with summarizing sticky notes, generating ideas, and even automatically organizing content. It’s like the digital equivalent of that one colleague who loves turning chaos into order.

FigJam AI: FigJam leans into being your brainstorming buddy. Think ideation prompts, group clustering suggestions, and contextual tips to help you keep momentum during your session.

On paper, they sound like they’ll save time and make workshops feel seamless. But how do they work in real scenarios?

The Tests: Putting AI to work

I ran three tests with each platform to compare their AI assistants in action.

1. Generating ideas:

I used the prompt: “Suggest ideas for a new team-building activity.”

Miro delivered 15 sticky notes with team-building activity suggestions. While the breadth was decent (think art workshop, charity volunteer day, and escape rooms), it felt like a brainstorming session you could get from a quick Google search.

Curious to test the depth, I gave it a follow-up prompt:

“Deep dive into the escape room challenge.”

The response? Miro provided a generic list of things to consider, like “theme,” “venue options,” and “team size.” While somewhat useful, it didn’t add much value beyond surface-level suggestions.

FigJam’s Take

FigJam took a slightly different route. Instead of listing a variety of activities, it structured a potential workshop, including an ice-breaker, collaborative challenges, and a reflection session. While this approach felt more actionable, it leaned heavily on process rather than creative ideation.

To match Miro’s test, I gave FigJam the same follow-up prompt:

“Focus on the escape room challenge.”

FigJam responded with a mind map that broke down the activity into key elements: Theme ideas (e.g., haunted house, pirate adventure), Puzzles (e.g., logic games, physical challenges), Team dynamics (e.g., roles, communication tips), Time constraints (e.g., countdown, pressure) and Room elements (e.g., props, lighting).

This additional layer of detail made FigJam’s response feel more tailored and creative, with practical suggestions that extended beyond just the initial idea.

Key takeaways

• Miro excels in generating a variety of ideas but struggles to go deeper without additional input.

• FigJam shines in creating a structured framework and offers more detailed follow-ups, particularly for planning activities like the escape room challenge.

Round 1 results: 🏆 FigJam takes the win!

In this round, FigJam’s ability to add meaningful depth and organise the information gave it the edge — especially for teams that want to move quickly from ideas to execution.

2. Organizing content:

I tested both Miro and FigJam by dropping 20 sticky notes into each board and asking their AI to organize them into clusters.

Miro handled this effortlessly, grouping the words into meaningful categories with ease. The user experience was straightforward — simple to navigate — and the results were spot on.

FigJam, on the other hand, struggled. While the UX was equally intuitive (you select items and use a dropdown to cluster), the output was far less reliable. For instance, it categorized my cat as furniture and my breakfast banana as a color!

As a bonus, I decided to push Miro further by testing it with a more challenging list of words. Once again, it performed impressively, grouping the words into accurate and meaningful clusters without much trouble.

For comparison, I fed the same list of words to ChatGPT to see how it would cluster them. Here's the result:

“Here’s a possible grouping for the challenging list:
- Natural Elements: Quartz, Ash, Seaweed, Lichen, Maple, Dandelion, Thorn
- Animals: Wolf, Penguin, Salmon, Scorpion, Bat
- Tools: Shovel, Wheelbarrow, Pulley, Wrench, Grinder
- Guidance/Reflection: Compass, Mirror, Lantern

These groupings reflect different overarching themes like nature, animals, tools, and abstract or symbolic items. However, some words like “Ash” or “Lantern” could fit multiple groups depending on context, making it less straightforward.”

Overall, Miro clearly outshines FigJam in clustering accuracy and reliability.

Round 2 results: 🏆 Miro comes out top!

3. Summarizing key takeaways:

Finally, I created a deliberately chaotic mock brainstorming session to simulate the typical messiness of workshop outputs — fragmented ideas, incomplete thoughts, and scattered priorities. The session was split into two sections: one for ideation (10 sticky notes) and another for next steps (10 sticky notes). I then asked each AI assistant to summarize the workshop.

Miro excelled at making sense of the chaos, transforming the output into a clear and structured document. It effectively grouped the ideation section into relevant sub-themes and organized the next steps, making the summary practical and easy to use.

FigJam, on the other hand, struggled to separate ideation from next steps, which made the summary harder to follow. While the addition of emojis was a fun touch, the lack of clear differentiation made the output less helpful.

Once again, Miro came out ahead, providing a well-organized and actionable summary that effectively categorized the workshop’s key points.

Round 3: 🏆 Miro wins (again)

4. Approaches to AI: pragmatic vs. playful

The contrast between Miro and FigJam’s AI assistants ultimately reflects their different problem-solving philosophies:

• Miro: Pragmatic and utilitarian. Miro’s AI is built for teams that value order and efficiency. It’s a dependable assistant for summarizing, organizing, and handling chaotic inputs with precision. Perfect for corporate workshops, project managers, or structured strategy sessions where clarity is non-negotiable.

• FigJam: Playful and creative. FigJam’s AI brings a more dynamic, brainstorming-friendly vibe to the table. Its strength lies in generating ideas, structuring frameworks, and offering fun, collaborative moments — ideal for design teams, startups, or anyone looking to inject personality into their workshops.

5. The verdict: which one should you Use?

And the winner is.. 🏆 Miro!

Although the UI for the AI assistant in both Miro and FigJam is nearly identical — simple, intuitive, and easy to use — Miro consistently outperformed FigJam when given the same tasks. From organizing content to summarizing chaotic inputs, Miro delivered more accurate, actionable results, making it the clear winner in the tests I conducted. That said, it’s worth noting that these tests were not exhaustive, and the results could vary depending on the specific use case or scenario.

Final thoughts: AI in whiteboards — game-changer or just hype?

Both Miro and FigJam are integrating AI in ways that speed up mundane tasks and enhance collaboration, but they’re far from perfect. A notable limitation I found is the inability to refine outputs or give specific feedback to the AI assistant. Without the flexibility to go back and forth, their potential is sometimes undercut by rigidity.

Still, these tools push the boundaries of how we collaborate, and it’s exciting to imagine how they’ll evolve to better anticipate user needs.

For now, I find myself using both — Miro when I need precision and FigJam when I want creativity. What about you? Have you tried their AI assistants yet? Which one do you think works best for your team?.

Here are some topics/websites I recommend going through

If you want to go further, I can’t recommend enough reading:

• How UX can be a key differentiator for AI

• Design as Thought: AI and the Future of Design

Any more? What’ve I missed. Let me know in the comments

Follow me on LinkedIn & Substack for more deep dives 🧜🏻‍♀️

I’ve spent years building products, writing about behavioral design, and nerding out over how tiny nudges—like badges in apps or the perfect push notification—can change our habits. It’s fascinating stuff, but let’s be real: sometimes it’s brilliant, and sometimes it’s just plain chaotic. (Looking at you, apps that hand out badges for opening them twice in a week…🙄)

I plan to share actionable insights to build for people, a biweekly, psychology-driven tips to help you create better products, understand users, and grow as a leader.

Let me know what you’d love to hear about—no idea is too nerdy, too niche, or too big. Thanks for subscribing and joining me on this ride. 🚀