PProduct Joi

Customer & Discovery

Problem definition / opportunity brief

Tight articulation of a specific problem worth solving.

What it is

A 1-pager that defines user, context, problem, why now, impact, and constraints. Forces the team to align on the problem before solutioning.

Concrete example

“Techs often arrive unprepared; 22% of jobs require return visits. Opportunity: improve pre-job information flow to reduce callbacks by 30%.”

When to use it

Before scoping any feature larger than a small fix. Whenever someone says “let's just build it.”

What Product Joi delivers

1-page opportunity brief: user, context, problem, evidence, why now, expected lift, success metric, constraints, out-of-scope.

  • Turnaround

    48 hours

  • Format

    Decision-first doc

  • Reviewed by

    Senior product lead

Sample deliverable

A real problem definition / opportunity brief Product Joi has shipped. This is the format and depth you'll receive.

Opportunity map · shipped 6 weeks ago

AI opportunity map — Dispatch workflow

Maps every AI opportunity in dispatch against impact, feasibility, and trust risk. Picks one to pilot and parks the rest with rationale.

Decision Pursue AI route-suggestion. Park crew-shift recommender for next half. Drop chatbot deflection entirely. The map We scored every dispatch workflow on three axes — impact on TTFV, ML feasibility, and trust risk if wrong — then plotted them. High impact / high feasibility / low trust risk: 1. Route-suggestion (winner). Telemetry shows 78% precision on hold-out set; reversible. High impact / medium feasibility / high trust risk: 2. Crew-shift recommender. Parking — wrong suggestion breaks payroll trust. Medium impact / low feasibility: 3. Predictive ETA. Park — data sparsity in current cohort. Low impact / any feasibility: 4. Chatbot deflection. Drop — prior effort showed no movement on activation, retention, or TTFV. Re-litigating wastes calendar time. Why route-suggestion wins - It hits the friction point in the funnel (initial route config is 41% of TTFV variance). - Wrong suggestions are obviously wrong, so trust damage is bounded. - We can pilot it on Wave 1 in 10 days; we'd need 8 weeks of new data to pilot anything else. Risks - We over-index on one workflow and ignore unrelated AI investments. Mitigation: revisit map at quarter end; cap pilot capacity at 1 active AI bet at a time. Next step Write pilot plan (separate deliverable). Stand up daily accept-rate signal. Set a quarterly review on this map.

Related artifacts in Customer & Discovery

Customer needs / insights report

Synthesis of qualitative and quantitative research.

User personas

Archetypes of key user segments.

Buyer personas

Archetypes of economic decision makers.

← Back to Customer & Discovery

Ready for a problem definition / opportunity brief in 48 hours?

Start subscriptionBook diagnostic