Concept to code: Transforming ideas into functional products with Kiro
This video is featured in the AI Conference playlist.
Summary
AI-assisted development is rewriting the contract. Kiro, an agentic IDE built around specification, puts the power to build in the hands of the person who already holds the intent. The design.md spec becomes the design output. Static mocks become shippable code. No relay, no translation, no slow erosion of meaning between rooms.
In this session, Rikki Teeters takes an idea from spec to functioning product live on stage. You will watch the spec take shape, see Kiro build against it, and feel the moment a concept becomes code in real time. Expect a working grasp of Kiro, a new frame for the spec as the most consequential artifact a designer makes, and one question worth carrying home. When execution is no longer someone else’s job, what will you build first?
Key Insights
-
•
Traditional design-to-engineering handoffs are effectively dead due to AI-driven direct code generation.
-
•
AI enables designers to ship functional products themselves, removing the need for separate coding phases.
-
•
The new builder team model blurs roles between design, engineering, and product management, with AI as a core team member.
-
•
Design.md is a new kind of design spec written to guide AI in building products, replacing fragmented mockups and annotations.
-
•
Spec-driven AI development leads to higher quality and fewer errors compared to vague prompt-based coding, reducing 'AI slop'.
-
•
About 80% of code merged to production at Anthropic is now written by AI (Claude), showing AI's growing role in development.
-
•
AI coding parity with humans was reached recently, and AI is expected to outperform humans in coding within a year.
-
•
UX professionals can act as directors of AI development, controlling product vision from design through to launch.
-
•
Tools like Kiro facilitate both vibe coding (conversational iterative prompts) and spec-driven coding for designers.
-
•
Including personas, user flows, and edge cases in the design.md spec helps AI build more human-centered and accessible products.
Notable Quotes
"More times than not, the product looks quite different than the vision we designed by the time it ships."
"The traditional handoff between design and engineering is officially dead."
"We’re able to use AI to create the vision and develop that vision into real working code and ship it ourselves."
"When anyone can make anything very fast and cheap, the role of UX is evolving."
"At Anthropic, 80% of the code merged to production is now written by Claude."
"AI writing code at parity with humans is a huge achievement, and it’s expected to surpass humans within a year."
"Design.md is not technical specs but a document that you can give to AI to instruct what it actually builds."
"AI slop means a lot of low quality AI-generated content with little human thinking and decision-making involved."
"If you can describe your idea clearly enough that someone else could build it, you can have AI build it for you."
"Stop handing off your vision and start shipping it yourself."
Or choose a question:
More Videos
"Some engineers build a council of agents that debate and generate responses, which can help with bias and accuracy."
Daniel KorczynskiWhy AI Is Bad at Research (and how to make it actually useful)
March 10, 2026
"We answer questions differently when we’re living life and doing activities than when we’re sitting down facing the camera."
Bas Raijmakers, PhD (RCA) Charley Scull Prabhas PokharelWhat Design Research can Learn from Documentary Filmmaking
March 11, 2022
"Belgian government services are still really designing with desktop use in mind, even though smartphones are more accessible."
Yalenka Mariën Marie MervaillieDesigning for Digital Inclusion in the Belgian Government
December 8, 2021
"Too often research lives in silos, making it hard to connect the dots across streams."
Jake BurghardtStop wasting research: Unlock more value from research insights
June 24, 2025
"Every innovation carries responsibility; we must tread carefully with AI in health."
Vincent BrathwaiteOpener: Past, Present, and Future—Closing the Racial Divide in Design Teams
October 22, 2020
"Communicating with precision and nuance to get good AI outputs is not second nature; it really takes practice."
Christian Crumlish Erika Flowers Benjamin Jackson Adekunle OduyeFrom tools to staff: What the next generation of agents means for the future of design
June 10, 2026
"90% of everything is crud."
Adrian HowardSturgeon’s Biases
September 25, 2024
"Automated tests only get you so far; manual QA by people aware of accessibility needs is essential."
Megan Clegg Michael Haggerty-Villa Alexis MorinSpace for Everyone: Reframing Accessibility Through a Wider Lens
June 10, 2021
"You have to find your champion as high up in the company as possible to push for research and equitable pay."
Mike DavidsonFireside Chat
March 11, 2022
