Show-iNstead of-Tell Part II: CANOPY
Another experiment. Another lesson the building taught me. Another tool for you to try.
Two weeks ago I shared Innovation Coach — a tool that runs a question through multiple expert lenses and reframes the problem before answering it. 150 people tried it! And the feedback was clear: the reframe was something no general-purpose LLM produced. But Gemini and others gave more actionable results. Fair.
CANOPY is the response. The next experiment. Still a vertical slice — not a full product, just enough architecture to learn whether the idea holds weight. I built it in a day using Claude as my IDE: chat to design to code to deployment.
The experiment worked in that it taught me something I wouldn’t have learned any other way.
What CANOPY Does
Instead of one question and one response, CANOPY is a four-step journey: set the stage, see what’s actually there, shape the decision, build the path. Multiple expert lenses — strategic foresight, behavioral intelligence, submerged value, narrative — run independently, then converge. The tool reframes your question and coaches you through creating a commissioning brief and a 90-day action plan.
A preloaded test case — fictional company, fictional problem, fictional context — lets you walk through the experience without bringing your own. You can also overwrite the test case with your own to really see (and feel) how it works.
What CANOPY Taught Me
I layered in the lab’s own behavioral intelligence engine — a thin version of CORTX — alongside methodologies from experts whose work shapes how I think: Rita McGrath’s strategic foresight, Daniel Aronson’s submerged value framework, and Angus Fletcher’s narrative intelligence. These are my interpretations of their published work — not the full methodologies, not endorsed. Rita, Daniel, and Angus each bring far more depth than this experiment can carry — seek them out for the real deal.
Claude as an IDE got me further than I expected. The main workflow — the four-step pipeline, the coaching pauses, the engagement modes — got to a place I liked. Claude Design produced something better than what I would have designed alone. I don’t code, so Claude Code and Replit did all the heavy lifting there.
But here’s what I lost.
On Friday — the day before this build — I sat in a design session with my partners Dan and Kent for one of the Lab’s expert products in development. What I built on Saturday over 8 hours isn’t in the same league as what they produced in 90 minutes — and you wouldn’t need an expert eye to see it.
And then the bigger loss. Privacy by design, observable and transparent AI, confidence scoring — requirements I know matter to us — fell out of my head entirely while I was designing the core workflow. I added some of it back at the end, but bolting things on after the fact undermines the whole discipline of diverge-converge-synthesize. I know that better than anyone. Building alone, even with excellent tools, I could only actively hold so much. With my team, we each hold different aspects of the work — and the whole moves forward together.
The Lesson Underneath the Lesson
This build demonstrated what I know to be true: the expert building alone, even with AI that extends her capability, loses signal that the expert building with complementary collaborators wouldn’t.
That’s not a failure of the tools. It’s our thesis.
The premise of Regenerous Labs is that isolation is the risk. Expert methodologies get stronger in relationship with other expert methodologies — sequenced, layered, pressure-tested. I’ve shared that before. I believe it. Then I built alone for a day and proved it from the inside.
The vertical slice did its job. CANOPY’s architecture — the pipeline, the independent lenses converging, the coaching stance — holds. What doesn’t hold is me doing this by myself. Not because I can’t build. Because what we’re building requires more eyes and minds than mine.
This is what Regenerous Labs is here to do — give experts the infrastructure to build in relationship instead of alone. CANOPY is the latest experiment. Try it. Tell me what I got wrong. Help me build it better.
Try CANOPY
There’s no storage, no database — everything stays local. We can’t see what you do. This is an experiment, and we’d love your feedback. Add a comment here or use the feedback tool in the bottom right corner of the app (click the Regenerous icon).
Connections to The Insider’s Guide to Innovation at Microsoft
Developing in the Open — The VS Code story: sharing unfinished work accelerates learning and builds trust faster than polished launches
Start With Who — Assembling complementary collaborators before defining solutions; the build revealed what happens when you skip this
Say It Ugly — The series as the practice; this piece as the latest demonstration
Behavior is the Barrier — The gap between knowing what matters and executing it under constraint
This piece was created with the help of AI — specifically Claude, Perplexity, and a team of expert personas built by Regenerous Labs. Direction, judgment, and final decisions by me. CANOPY itself was built using Claude as an integrated development environment. Say it ugly, build it better. Onward!
