How we choose what to build, and how we build it.
Every project asks three questions: do people need it, can we build it well, and will it help the business over time? The workflow gives teams a shared way to answer those questions and move forward with confidence.
For a long time, building was the expensive part. We rationed it. Roadmaps, intake rituals, and most of our org chart were designed around that one constraint.
AI has made many implementation tasks faster and easier. That makes the human work more important: framing the problem well, choosing the right bet, and reading the result honestly. Use the time AI gives back to make those decisions stronger. That is where the leverage lives now.
Use the workflow on your next project
You do not need to run every stage from the top. Pick up the stage your team is actually in, name the weakest lens, and decide what evidence would help you move forward.
- Current stage: Where is the work today?
- Weakest lens: Desirability, feasibility, or viability?
- Next evidence: What would help the team decide with more confidence?
Three lenses, one overlap
Innovation becomes possible in the overlap where desirability, feasibility, and viability all line up. The workflow keeps all three in view, even when one feels more urgent than the others. Click a lens to see what it asks of you.
Design thinking lenses
Innovation
Where all three meet, and energy compounds.
Questions to ask
- Have we proved each lens with evidence, not opinion?
- Which lens is the weakest right now, and how would we know it has moved?
- What is the smallest thing we could ship to test all three at once?
- Is the team set up to keep all three lenses healthy after launch?
How to investigate
- Cross-functional reviews with all three lenses in the room
- Pre-mortems framed lens-by-lens
- Bets, not commitments. Write down what would make you change your mind
- Shared metrics that respect all three
The default failure mode is letting one lens dominate because it has the clearest owner or the loudest advocate. The workflow keeps all three lenses in the conversation.
Why the Venn matters more now
When building was expensive, teams often protected build time by shortening the thinking time. Just start coding could feel practical. AI makes that pull even stronger: the prompt is right there, and a working prototype can be one paragraph away. The risk is moving faster than the team’s understanding.
The work that pays off has moved upstream. A sharper problem statement can save weeks of refining a direction that should have changed earlier. A detailed spec, written before any code, turns AI from a guesser into a builder. Another hour with a real user can make the whole brief better.
Think first. Write the spec first. Then build with speed. The workflow helps the time AI gives you back land on the work that actually matters.
Six stages, three modes
The work groups into three modes (Understand, Explore, Materialize) and six stages within them. The arrows go forward, but the process loops back often. Testing sends you back to ideate. Implementation reveals what you missed in empathy. Click a stage to see what it looks like in practice.
The six stages of the workflow
Empathize
See the problem through the eyes of the people who have it.
What you do
- Sit with users in their context, not yours
- Observe what they actually do, not what they say they do
- Hold back on solutions. The moment you propose one, you stop listening
What you produce
- Interview notes and quotes that you can point to
- Journey or service maps of the current state
- A clear list of pains, gains, and workarounds
A felt sense of who you are designing for and where the friction lives.
Open a stage
Each stage has its own page with the questions to ask, the artifacts to expect, and how AI fits into that specific moment of the work.
- 01 · UnderstandEmpathize
See the problem through the eyes of the people who have it.
Open - 02 · UnderstandDefine
Name the problem clearly enough that the best options are easier to see.
Open - 03 · ExploreIdeate
Generate many options, defer judgment, then converge.
Open - 04 · ExplorePrototype
Make it real enough to learn from, no more.
Open - 05 · MaterializeTest
Watch what happens. Let evidence guide the next move.
Open - 06 · MaterializeImplement
Ship it, instrument it, keep learning.
Open
Where this lives in the system
The workflow does not stand alone. Lean on these pages as you move through it.
- Design Principles
The values that bias the workflow. What we choose when we have to choose.
- Anti-Patterns
The shapes of work that look like progress but are not. Read before any review.
- Voice & Tone
The voice we use when the product talks to a human. Shapes the empathy stage and every page after.
- Decision Log
Every meaningful trade-off the workflow produced, written down so the next team inherits the reasoning.
Start where the energy is.
You do not have to begin at Empathize. Start at the stage you are actually in. The workflow keeps all three lenses in view from wherever you pick it up, and it points the time AI gives you back toward the parts of the work that matter most.