The Problem to Product Workflow
A decision and evidence system for turning uncertain problems into validated product bets. Ten connected decision stages that loop as evidence changes, with artifacts pulled by decisions — not pushed by process.
This is not a phase gate. Discovery and delivery run continuously: a prototype in stage 7 can break an assumption and send the work back to stage 1, and a learning review in stage 10 feeds the next bet.
What the Workflow Produces
Four working outputs come out of any project that runs through the system, regardless of domain or stage. Nobody hires a PM to create documents — the output is better product decisions under uncertainty.
Product Clarity
A defined problem, user, segment, job, and why-now context that everyone can point to.
Tested Bets
An opportunity hypothesis with its riskiest assumptions named and tested cheaply before heavy commitment.
Execution Readiness
PRDs, user stories, acceptance criteria, scenarios, and prototype direction a team can build from.
Learning Evidence
Launch monitoring, post-launch reviews, PMF checkpoints, and the next decision, recorded honestly.
The Questions I Answer Before, During, and After Building
Every stage exists to answer a specific product question. The map below shows which question belongs where — and evidence can reopen any of them.
| Stage | Product Question | Workflow Step |
|---|---|---|
| Problem | What real user problem are we solving, and why now? | S01 · Problem Discovery |
| User | Who exactly is this for, and which segment are we deliberately not solving for? | S02 · User & Segment |
| Context | How do users solve this today, and what are the alternatives, substitutes, and constraints? | S03 · Context & Alternatives |
| Bet | What is the product bet, the success outcome, and the riskiest assumption underneath it? | S04 · Product Bet |
| Validation | Which assumptions could kill this, and what is the cheapest test for each? | S05 · Assumption & Risk Testing |
| Scope | What is the smallest useful version, and what are the visible trade-offs? | S06 · MVP Scope |
| Experience | How does the user actually accomplish the job, including edge cases and error states? | S07 · Experience & Prototype |
| Definition | Can a team build this from the document, end to end? | S08 · Build-Ready Definition |
| Launch | How do we ship it, monitor it, and capture feedback? | S09 · Delivery & Launch |
| Learning | What did we learn, and what is the next decision: iterate, pivot, pause, or sunset? | S10 · Learning & Next Decision |
Stage-by-Stage
Open any stage to see its purpose, decision focus, the tools I reach for, and how it shows up in a live product. Arrows between stages point both ways: evidence decides the direction.
The Artifact Toolkit
Artifacts are pulled by decisions, not pushed by the workflow. The question comes first: what evidence do I need, and what is the lightest artifact that helps? A simple project might touch five of these; a complex one, twenty. Never all of them.
Framework Index
The frameworks I lean on, grouped by the product decision they support. A toolkit, not a checklist: the decision picks the framework, never the other way around.
One Core System, Optional Adapters
The core workflow is framework-agnostic. Delivery-method specifics plug in as adapters when the environment calls for them — a startup never sees a PI Objective, and a SAFe shop never loses the discovery loop.
For organizations running the Scaled Agile Framework.
For products with an AI or ML component.
For early-stage, evidence-starved environments.
How I Label Evidence
Every project gets an honest evidence label. The label tells you what stage the proof is at, not how impressive the headline sounds. This is the single maturity scale used everywhere on this site.
How I Use AI-Led Development in the Workflow
AI-led development supports the workflow at specific points: research synthesis, documentation drafts, UI exploration, prototype scaffolding, internal tools, and technical probes. It speeds up the parts that benefit from speed.
Final product judgment stays human-led. Problem framing, user understanding, scope decisions, trade-off calls, prioritization, and learning reviews are reasoned through and signed off by me, not by a tool. AI-generated material is treated as a draft that needs review, testing, and clear disclosure.
See the System in Live Products
The system is most useful when you can see what it ships. SimpliLEAD, Catalyst Solution Services, and HiMirrorly are live products in production today — each one built through these same ten decision stages.