The Double Diamond Meets AI
How AI changes the economics of our favorite design framework.
Reading time: 12 minutes
For twenty years, the Double Diamond has been our north star. Discover, Define, Develop, Deliver. It's elegant. It's teachable. It's on every job description and in every design org's process docs.
And it made sense…when it was created.
In 2005, when the Design Council introduced the Double Diamond, making things was expensive. Every prototype meant hours of design work, developer involvement, and careful resource allocation. User testing required recruiting participants, booking facilities, and waiting for results.
So we front-loaded the "Discover" and "Define" phases. We researched exhaustively. We synthesized meticulously. We aligned stakeholders. We wrote briefs and principles, and strategy docs. All to reduce risk before we commit to the expensive part: making.
That logic just broke.
AI has collapsed the cost of making to nearly zero.
What used to take you three days—mocking up five interface variations, writing contextual copy, generating a working prototype—now takes thirty minutes. Testing that requires scheduling and logistics can happen continuously. Synthesis that meant days of affinity mapping now happens in real-time.
When making was expensive, the Double Diamond was risk management. When making is cheap, the Double Diamond is a waste.
Let me show you the actual cost.
The Double Diamond in Practice: A Real Project
Your team needs to redesign a core user dashboard. Here's who's involved:
| Role | Salary* | Hourly* (with 30% overhead) |
|---|---|---|
| Product Manager | $130,000 | $81 |
| UX Designer | $115,000 | $72 |
| UX Researcher | $115,000 | $72 |
| Front-end Developer (Senior) | $140,000 | $88 |
| Front-end Developer (Mid) | $120,000 | $75 |
Discover (Weeks 1-2)
Researcher conducts 15 user interviews (80 hrs × $72 = $5,760)
PM observes half the sessions (40 hrs × $81 = $3,240)
You observe select sessions (20 hrs × $72 = $1,440)
Define (Week 3-4)
Full team synthesis workshop (5 people × 8 hrs)
PM: 8 hrs × $81 = $648
You: 8 hrs × $72 = $576
Researcher: 8 hrs × $72 = $576
Senior Dev: 8 hrs × $88 = $704
Mid Dev: 8 hrs × $75 = $600
Researcher creates insights deck (30 hrs × $72 = $2,160)
You create journey maps and personas (20 hrs × $72 = $1,440)
PM writes strategy doc (25 hrs × $81 = $2,025)
You create design principles (15 hrs × $72 = $1,080)
Stakeholder alignment meetings (5 people × 4 hrs)
PM: 4 hrs × $81 = $324
You: 4 hrs × $72 = $288
Researcher: 4 hrs × $72 = $288
Senior Dev: 4 hrs × $88 = $352
Mid Dev: 4 hrs × $75 = $300
Develop (Weeks 5-9)
You explore three directions (60 hrs × $72 = $4,320)
Weekly reviews (5 people × 4 hrs × 2 weeks)
PM: 8 hrs × $81 = $648
You: 8 hrs × $72 = $576
Researcher: 8 hrs × $72 = $576
Senior Dev: 8 hrs × $88 = $704
Mid Dev: 8 hrs × $75 = $600
PM prepares stakeholder presentation (15 hrs × $81 = $1,215)
You prepare stakeholder presentation (15 hrs × $72 = $1,080)
You revise based on feedback (30 hrs × $72 = $2,160)
You refine chosen direction to high-fidelity (80 hrs × $72 = $5,760)
Senior Dev reviews for feasibility (20 hrs × $88 = $1,760)
Mid Dev reviews for feasibility (10 hrs × $75 = $750)
Deliver (Weeks 10-12)
Senior Dev builds interactive prototype (60 hrs × $88 = $5,280)
Mid Dev builds interactive prototype (60 hrs × $75 = $4,500)
You provide assets and specs (20 hrs × $72 = $1,440)
Researcher recruits and runs 10 usability tests (60 hrs × $72 = $4,320)
PM observes sessions (20 hrs × $81 = $1,620)
You observe sessions (20 hrs × $72 = $1,440)
Researcher synthesizes findings (20 hrs × $72 = $1,440)
You iterate based on findings (40 hrs × $72 = $2,880)
Double Diamond Project Breakdown
It’s week 12. You've followed the process. You've done the research, the synthesis, the alignment, the iteration. You're finally testing with users.
And they tell you the dashboard isn't what they need. They need notifications. 😱
Twelve weeks. $64k. And you're starting over.
AI-Enabled Experimentation: Same Team, Different Process
What if you didn't wait until week 12 to learn what users actually need? What if you tried to make and then test in shorter and shorter loops?
AI-Enabled Experimentation allows us to speed up the learning, front-load the risk, and prioritize getting prototypes in the hands of users as soon as possible. It’s a continual loop of
Here’s what it might look like in a typical project:
Day 1: First Hypothesis
You use AI to generate 3 crude dashboard concepts (4 hrs × $72 = $288)
You create a quick functional prototype (2 hrs × $72 = $144)
Day 2-3: First Learning
Researcher runs quick remote tests with 5 users (10 hrs × $72 = $720)
PM watches live, takes notes (6 hrs × $81 = $486)
You watch live, take notes (6 hrs × $72 = $432)
Team synthesis: what did we learn? (3 people × 2 hrs)
PM: 2 hrs × $81 = $162
You: 2 hrs × $72 = $144
Researcher: 2 hrs × $72 = $144
Day 4-5: First Pivot
You create version 2 based on findings (6 hrs × $72 = $432)
Researcher tests with 5 different users (10 hrs × $72 = $720)
Discovery: users don't want a dashboard—they want notifications
Week 2: Follow the Evidence
You prototype notification system (8 hrs × $72 = $576)
Researcher tests with users (10 hrs × $72 = $720)
You iterate on tone and timing (6 hrs × $72 = $432)
Researcher tests again (10 hrs × $72 = $720)
Week 3: Refine What Works
You add complexity based on what's working (12 hrs × $72 = $864)
Senior Dev assesses technical feasibility (8 hrs × $88 = $704)
Mid Dev assesses technical feasibility (8 hrs × $75 = $600)
Researcher runs A/B test with two approaches (12 hrs × $72 = $864)
Team synthesizes patterns (5 people × 2 hrs)
PM: 2 hrs × $81 = $162
You: 2 hrs × $72 = $144
Researcher: 2 hrs × $72 = $144
Senior Dev: 2 hrs × $88 = $176
Mid Dev: 2 hrs × $75 = $150
Week 4: Ship with Confidence
You create detailed specs (16 hrs × $72 = $1,152)
Senior Dev builds production-ready version (30 hrs × $88 = $2,640)
Mid Dev builds production-ready version (30 hrs × $75 = $2,250)
Researcher runs final validation testing (8 hrs × $72 = $576)
AI Experimentation Project Breakdown
Comparing the Double Diamond vs. AI Experimentation
You saved $47,724 and 8 weeks. In this example, you could run an AI Experimentation loop in one-third the time it takes you to use the Double Diamond.
Project Comparison
But here's what really matters.
In the Double Diamond, you discovered the concept was wrong at week 12. Starting over means another 6-8 weeks and $30-50k.
In the Experimentation Loop, you discovered it on day 5. The cost of that pivot? Six hours of your time. About $430.
Double Diamond
AI Experimentation
This Isn't Abandoning Research
You might be thinking: "But the research matters. The synthesis matters. You can't just skip to making."
You're right. Research matters. Understanding users matters. Strategic thinking matters.
But here's what the Double Diamond gets wrong: it assumes these things happen in phases. That you discover, then define, then develop, then deliver. That you finish one diamond before starting the next.
AI-powered experimentation loops don’t skip research; it is research. You're still talking to users. You're still synthesizing patterns. You're still making strategic decisions. You're just doing it continuously, in response to real evidence, instead of front-loading it based on predictions.
The fastest way to understand user needs isn't a two-week discovery phase. It's putting something concrete in front of them and watching what happens.
The best insights don't come from journey maps and personas. They come from observing real reactions to real prototypes.
The Skills That Matter Now
This shift changes what it means to be a senior product designer.
The old model rewarded designers who could synthesize research into elegant frameworks, create comprehensive documentation, and defend their decisions in stakeholder reviews. Those skills still matter, but they're no longer differentiating.
The new model rewards designers who can:
Prototype faster than they can plan. When you can make five variations in an hour, you stop debating which one is "right" and start testing to find out. Just do it.
Design for learning, not a handover. The goal of early work isn't a polished artifact; it's a clear hypothesis you can validate or invalidate. That’s how your researcher & PM thinks, and it’s how you should think, too!
Synthesize in real-time. Instead of waiting until you have "enough" research, you're continuously integrating what you learn into what you make. Even a few minutes of synthesis can add up if you do it every day.
Let go of ownership. When you're iterating daily based on user feedback, there's no "your design" to defend. There's only what works and what doesn't. Doesn’t that make you feel free?
Make the pivot cheap. The real skill isn't avoiding wrong directions because that is impossible. The real skill is structuring your work so that wrong directions cost hours, not weeks.
Find creative ways to test. Once you start experimentation, you realize that finding users to test things is the bottleneck. Finding creative ways to get data is a skill all on its own. And one we teach at the Fountain Institute.
Generate data to improve decisions. Don’t guess or rely on your instincts when you can supplement every decision with data from your users. If you can do that, you will make your PM’s day.
The Double Diamond Was Right for Its Time
The Design Council wasn't wrong in 2005. When making was expensive, front-loading research and strategy made economic sense. The Double Diamond was an elegant solution to a real constraint.
But constraints change. And when they do, processes need to change with them.
AI hasn't just made design faster. It has inverted the economics on which the Double Diamond was built. The expensive part is no longer making…it's waiting. Waiting to test your assumptions. Waiting to learn what users actually need. Waiting to discover you were wrong.
The designers who thrive in this new reality won't be the ones clinging to familiar frameworks. They'll be the ones who recognize that the tools have changed, the economics have changed, and the process needs to change, too.
The Double Diamond served us well.
It's time to move on.
*Salary sources (November 2025): Glassdoor, Indeed, Built In, Salary.com.
*Hourly rates: include 30% overhead for benefits, taxes, and employment costs.
Learn More about Experimentation
Watch How to Design Product Experiments a 60-minute masterclass on testing assumptions and de-risking work with experiments
Read The Designer’s Guide to Testing and Experiments and learn the experimental mindset needed for assumption testing
Read What is Data-Driven Product Design? and learn about the way to approach data and decision-making when designing experiments
Download the Experiment Cards to turn solution ideas into product experiments
Watch this talk from an innovation designer, at the Fountain Institute meetup to see how she tests innovative ideas
What’s your take? Does the Double Diamond still have some use in the age of AI?
