Blog
Notes on building AI products.
Playbooks and essays for early founders — validating an AI idea, scoping an MVP that ships, and designing AI people actually trust.
Write the BRD before you write the PRD
A BRD is the business clarity layer most rushed product builds skip. Here is how founders can use one to align outcomes, stakeholders, scope, risks, and success metrics before product work begins.
Write a PRD designers, engineers, and AI agents can actually use
Most PRDs are either too vague or too bloated. A useful PRD defines users, flows, states, data, priorities, acceptance criteria, and boundaries clearly enough for people and AI agents to build from.
A design system is product infrastructure, not decoration
A founder-ready design system defines tokens, components, patterns, content rules, accessibility, governance, and engineering handoff so product quality survives speed.
Do not start user interviews without a research plan
A user research plan turns vague discovery into decisions. It defines what to learn, who to study, which methods to use, and how findings will change the product.
Your AI coding agent needs a brief, not a vibe
AI coding agents work better when they receive product context, repo rules, constraints, acceptance criteria, and verification commands. Here is the brief structure founders should use.
A good SOW prevents scope drift before it starts
A Statement of Work should turn approved scope into deliverables, responsibilities, assumptions, timeline, acceptance criteria, and commercial boundaries.
A research repository keeps customer learning alive
Research loses value when findings scatter across docs, calls, and slide decks. A lightweight research repository keeps evidence, decisions, and product opportunities reusable.
Information architecture should happen before interface design
Information architecture gives a product its structure before the pixels arrive. It clarifies navigation, page relationships, content groups, objects, and user paths.
Technical architecture is a product decision
Technical architecture is not only engineering planning. It defines the system boundaries, data model, integrations, security, scalability, and tradeoffs that shape the product.
How to write personas that do not feel fake
Useful personas are not fictional biographies. They capture jobs, pains, triggers, objections, context, and product needs that help teams make sharper decisions.
A journey map turns user pain into product opportunity
A journey map helps founders see stages, actions, emotions, touchpoints, pain points, and opportunities before deciding what the product should improve.
Write the acceptance plan before launch panic begins
A QA and acceptance plan defines what must be tested, accepted, rejected, reviewed, and signed off before a product release goes live.
A launch checklist is a cross-functional document
A useful launch checklist coordinates product, engineering, analytics, support, legal, content, and go-to-market readiness before release.
Why every AI app looks the same — and what it costs you
AI made it trivial to ship a working product. It also made everything look identical. Here's why AI apps have collapsed into the same interface — and why deliberate design is now the differentiator, not the decoration.
AI is not your product. The experience is.
Founders keep pitching the model. Users only ever meet the experience. Why the intelligence is the cheap part, and the trust, clarity and judgment around it are what actually make an AI product succeed.
Design and engineering shouldn't be a handoff
In AI products, the design and the intelligence are the same decision — so splitting them across a wall produces generic, brittle results. Why the strongest AI teams fuse product design and engineering instead of passing work between them.
How to write a one-page PRD for your AI product
A one-page PRD forces the clarity most AI ideas lack. Here's the exact structure — problem, users, core loop, MVP scope, and risks — and how to write one a team can actually build from.
How to validate an AI product idea before you build
A practical playbook for early founders: how to tell whether your AI idea is worth building, the cheapest ways to test it, and the assumptions that quietly kill startups.
Sizing up the competition: how to find the wedge only you can own
'No competitors' is a red flag, not a green light. Here's how to map the AI landscape, read what users hate about existing tools, and find the narrow wedge you can actually win.
Finding the core loop: how to scope an AI MVP that ships
Most AI MVPs try to do everything and ship nothing. Here's how to find the single core loop that proves your product — and scope a build you can launch in weeks, not quarters.
Is your AI idea ready to build? A founder's readiness checklist
Before you spend months building, run your AI idea through this readiness checklist — problem clarity, the core loop, where AI fits, and the proof you still owe yourself.
Designing AI you can trust: patterns for control and transparency
Most AI features fail on experience, not the model. Here are the design patterns — control, transparency, and graceful uncertainty — that make an AI feature people actually trust.
The one assumption that can kill your startup — and how to test it this week
Every idea rests on a stack of assumptions, and usually one is fatal. Here's how to find your riskiest assumption and design a cheap experiment that proves or kills it fast.
Inline, not a tab: where AI actually belongs in your product
Bolting a chatbot onto your app is the easy path — and the reason most AI features go unused. The case for inline, in-context AI, and how to design it.
More on the way. Want to know when the next piece lands? Say hi and we'll keep you posted.
Get in touch →
