Does AI let designers and product managers independently prototype and iterate on products without relying on engineering?

Severin Hacker · May 19, 2025

A working prototype is far more persuasive than a design for winning internal buy-in on a new product, and AI tools now let non-engineers produce one

A PM and designer with little coding experience built the chess prototype with Cursor and other AI tools; a Figma design would probably have been rejected, but playing the prototype convinced everyone it belonged in the app

30:40 20VC: Duolingo Co-Founder on Why $3M is Harder than $100M to Raise | Why You Should Always Take Tier 1 VCs Even at Worse Terms | Why Europe Can't Win Unless the US Screws Up | How AI Impacts the Future of Work and Education with Severin Hacker

Harry Stebbings · Aug 18, 2025

These tools enable a democratization of idea generation inside companies, letting non-traditional builders like designers create shipped product features

He cites two Duolingo designers who built the first iteration of chess in Duolingo with such a tool

25:50 20VC: Lovable CEO Anton Osika on $120M in ARR in 7 Months | The Honest Truth About Defensibility and Unit Economics for AI Startups | The State of Foundation Models: Long Grok, Short OpenAI, Why | Replit vs Lovable vs Bolt: What Happens

Jason James · Aug 29, 2025

AI tooling has dramatically changed the prototyping stage: designers can now articulate intent in tools like Magic Patterns and produce an interactive prototype the whole team can see quickly

Their designers use Magic Patterns to express what they want and engineers can then vibe code from it

Scope: change observed within the last year

5:46 20Product: Why Most CPOs are Bad | Why You Do Not Need PMs in a World of AI | Why the Design Stage is Dead and How to Use Vibe Coding to Replace It | The Three Roles All Founders End Up Firing on Repeat with Jason James @ Tezi

Patrick Forquer · May 11, 2026

Non-engineers are now building working software themselves, and vibe-coded prototypes are an effective way to communicate product requirements

A funds attorney who couldn't articulate requirements to the product team vibe coded a version over a weekend, which was highly helpful in conveying the product vision

Scope: example drawn from Legora's legal engineers

60:49 20VC: Inside Legora: $100M ARR in 18 Months | Jude Law Generated $50M in Sales Pipeline: The Economics Broken Down | Competing Against Harvey, the 800 Pound Gorilla | Why Legora is Undervalued at $5.5BN with Patrick Forquer, CRO @ Legora

Cem Kansu · Jun 20, 2025

A PM and designer with no engineers can vibe-code a prototype good enough to validate an entirely new product line before any team is staffed.

On chess, the PM vibe-coded a full lesson experience and the designer made real designs; testing showed it was already a good Duolingo lesson, so they kept iterating on the prototype and only then staffed engineers — idea to launch in eight months.

Scope: chess was initially seen as a weird fit for Duolingo

11:56 20Product: How Duolingo Build Product 10x Faster with AI | Duolingo's Biggest Lessons on Paywalls, Push Notifications and In-App Purchases | Why Small Teams are the Future of Product | Why PMs Will Become Extinct with Cem Kansu, CPO @ Duolingo

Fernando Fanton · Jul 27, 2025

AI prototyping tools have made the search for product market fit meaningfully easier

Many companies used to fail just at putting something meaningful in front of a user to learn from the reaction; at Property Finder they now prototype and test with users in a day

13:44 20Product: Should We Kill the PM Role Entirely | How Does Product Design Change Most in a World of AI | How Do Monzo Build Product Today: What Works, What Does Not | Why Most Product Sucks and What Makes Truly Great Product

Aatish Nayak · Apr 11, 2025

Teams should prototype before writing the PRD

You can't know how to build the product or what the exact UX should be — especially with AI, where new interaction patterns are still being established — until you and internal customers actually play with it; the prework saves cycles later and yields a better outcome

27:12 Teams should prototype before writing the prd especially for new ai uxes

Aatish Nayak · Apr 11, 2025

For non-engineer prototypers, agent products that handle deployment like Replit agent mode and Lovable are more attractive than IDE-based tools like Cursor or Windsurf

He doesn't want to deal with deployment, front end, back end and servers, and Replit plans the build, lets you approve features and shows the code as it builds

46:13 Agent based deployment tools beat ide based tools for non engineer prototypers

Varun Mohan · Jun 2, 2025

For rapid prototyping, teams can skip the laborious design stage and go straight to mocking up designs consistent with company standards in code

If an app can be built very fast internally, going to Figma first isn't worth it; only work core to the product experience needs a full design pass

35:00 Skip formal design tools and mock up designs directly in code for speed

Scott Belsky · Dec 20, 2023

AI will make products meaningfully better by letting teams generate and evaluate multiple interface alternatives, so good solutions are reached in fewer cycles

Today exploring three or four interface approaches and abandoning dead ends takes a lot of time; AI can suggest variations you can eyeball or A/B test immediately, on top of code and bug-testing productivity gains

3:52 Ai generated alternatives shrink design iteration cycles

Gustav Söderström · Dec 20, 2023

AI is democratizing product work: designers and PMs can now arrive with a prompt and generate a working prototype instead of having to convince engineering to build one

21:42 Ai lets designers and pms prototype without engineering buy in

Scott Belsky · Dec 20, 2023

When core teams expose model capabilities as APIs, designers can be directly empowered to explore interfaces — and their job now extends to persona, intonation and conversational inflection rather than just visual design

Adobe ran early Firefly development as a design-driven exercise inside the design team to figure out interfaces and product integration

22:00 Exposing model apis empowers designers to explore interfaces directly