← Portfolio

DDM: Discovery & design mapping tool (DDM)

Autumn 2025 – present · Product & UX Designer, Developer

Interactive case Markdown export Full portfolio (llms-full.txt)

Outcome

DDM is a discovery and design mapping tool for product teams using generative AI. It connects outcome-led discovery via an Opportunity Solution Tree, from a chosen outcome through opportunities grounded in research and feedback to solutions and experiments, with structured design specs and markdown export so shared evidence and requirements exist before prototyping. Built from consulting experience and the friction of AI-assisted design work.

Interactive demo

Context

Across consulting engagements, two patterns kept repeating. It is hard to get approval to do research. When it is done, no one except the person who ran it cares about the raw material. Stakeholders and team members want crisp insights and clear direction, not transcripts. Designers often run their own research because they want control and credit for the full journey from research to feature. In practice, that means prior research gets ignored or duplicated.

Generative AI added a third layer. Prototyping with AI is fast, but when several people prompt from the same brief independently, the outputs diverge. Scope creep follows. There was no shared frame for mapping discovery evidence or for writing design requirements before people opened Figma Make, v0, or Copilot.

I have been developing this concept in parallel: as a side project in my own repo, and as a separate tool for a client with specific security requirements. It combines methods I have used as a consultant with what product design now demands when AI is in the loop.

Challenge

On the discovery side, research data and feedback pile up without a repeatable way to turn them into prioritised opportunities tied to a chosen outcome.

Teams working on AI-native product design lack a collaborative planning layer. Briefs still arrive fragmented: Jira tickets, PowerPoints, PDFs, screenshots, sometimes in different languages. Decisions get made from intuition or early conviction, and gaps show up later.

On the design side, generative tools amplify the problem. Same goal, different prototypes. Without a common spec, variations drift and AI-influenced scope expands quietly.

Three things needed to hold:
- Discovery mapping grounded in evidence, with AI helping analysis rather than replacing judgment.
- Structured design specs before AI prototyping, similar in spirit to spec-it but for design.
- Markdown export so the same context works across Figma Make, Spark, Copilot, and other tools.

Resources

Since autumn 2025, I have mapped the discovery-to-spec flow in FigJam and built the tool in parallel, iterating on structure and handoffs and rebuilding parts until the whole made sense. The board below uses made-up demo content to illustrate the model, not real client research or data.

Approach

Discovery mapping follows an Opportunity Solution Tree. The team picks an outcome to focus on. DDM uses AI to analyse existing qualitative research, quantitative signals, and feedback, then proposes opportunities the team can rank by business value and feasibility. Solutions and experiments stay human-led. I use AI as a sounding board for experiments, not as the ideation engine, because it lacks full context and can flatten creative perspective.

Design spec mapping. The second mode walks through problem, assumptions, edge cases, open questions, scope versions, and design references in a fixed structure. Paste from Jira, meetings, or PDFs; AI fills fields, but the sections stay consistent. Export produces a markdown AI brief: shared context for prototyping tools so different designers and different AI outputs can be compared against the same criteria.

I designed and built the tool solo. It exists in two completely separate environments: one repo for a multinational B2B client, one I maintain for myself. The client instance runs under their GitHub and Vercel accounts, with customisations for their security and workflow requirements; development there is limited to VS Code. My own repo deploys through my GitHub and Vercel accounts, where I work in Cursor, my tool of choice. Both are in active use: the client project for team discovery and design mapping, mine for day-to-day design work and ongoing iteration.

Impact

DDM is in use on a multinational B2B product project and in my own workflow. The practical value: a single place to map discovery from research through an opportunity tree to experiments, and a spec format that travels as markdown into AI design tools. Less ad hoc doc sprawl, clearer handoffs between discovery and prototyping.

Live embeds and FigJam boards use demo content; the tool itself runs on real engagements with client-specific data.

Reflections

The tool grew from consulting experiences. Keeping a side project version separate from a client build with security requirements let me iterate on the method without blurring ownership.

The Opportunity Solution Tree for discovery and markdown export for specs were the decisions that stuck. The open question is how much structure teammates want when AI makes it easy to skip straight to a prototype. I will keep refining DDM as a tool I can bring into engagements where discovery and design mapping need a shared frame.