Too many plausible use cases
The team needs one opportunity connected to a user, a business result, and evidence that can change a decision.
Five-day AI Product Direction Sprint
Know what deserves to be built—before the build gets expensive.
For founder-led SaaS teams with an important AI opportunity, competing assumptions, and a real decision to make. In five focused days, we narrow the release, test its largest uncertainty, and make the next commitment explicit.
Five working days · Fixed US$3,000 · Complete on its own
When the sprint fits
The opportunity feels real. The responsible release does not—yet.
This is the point between “we should do something with AI” and committing engineering time, budget, or organizational attention.
A promising idea with a critical unknown
The largest risk may be the experience, AI behavior, data, integration, or the workflow around the technology.
A team waiting for a build boundary
Product, design, and engineering need the same definition of what is in, out, later, and good enough to release.
The right proof for the risk
A prototype is common—but it is not automatically the answer.
The fit call and written scope identify one primary evidence artifact. That keeps five days focused on the uncertainty that can actually change the build decision.
Clickable prototype
Use when the core question concerns workflow, usability, comprehension, or stakeholder alignment.
Technical spike
A bounded experiment for AI behavior, data, architecture, or integration feasibility—not production software.
Workflow model
Use when the critical failure can travel across users, teams, operational steps, and connected systems.
Scope boundary: the sprint does not promise both a complete UX prototype and a technical proof of concept. It promises the evidence needed for the most important decision.
Five focused days
Each decision narrows the next.
Three client conversations anchor the week: kickoff, evidence review, and final decision. Analysis and production continue between them.
- Day 01
Understand
Align the user, business objective, constraints, evidence, assumptions, and decision.
Opportunity and evidence brief - Day 02
Choose
Prioritize one use case and define its smallest valuable release.
User journey and release boundary - Day 03
Test
Produce the agreed prototype, feasibility spike, or workflow model.
Primary evidence artifact - Day 04
Make buildable
Define technical direction, AI authority, evaluation, integrations, risks, and acceptance.
Technical and evaluation direction - Day 05
Decide
Synthesize the evidence and recommend build, narrow, buy, defer, or stop.
Decision package and walkthrough
What you receive
Enough definition to begin responsibly—or choose not to.
Every artifact is editable and designed for handoff. The sprint is useful whether your team builds independently, works with Gabriel, buys an existing solution, or stops.
- Product opportunity and decision brief
- Defined user, job, and intended business outcome
- Prioritized use case and critical journey
- Explicit in, out, and later boundaries
- One primary evidence artifact
- AI evaluation plan or conventional acceptance plan
- Technical direction and major integration decisions
- Assumptions, dependencies, risks, and unresolved evidence
- Success measures and release acceptance conditions
- 90-day implementation roadmap
- Preliminary delivery phases and estimate range
- Recorded walkthrough and editable handover
Clear boundaries
Five days only works when one decision leads.
Not a production build
A prototype or spike provides decision evidence. Production implementation, deployment, and ongoing support are separate.
Not open-ended discovery
Existing research and evidence are reviewed. Recruiting and running a new research program requires separate scope.
Not automatic presales
The final recommendation may be to use your team, another provider, an existing tool, a smaller release—or no build at all.
Choose the responsible first step
Bring one opportunity and the decision that is stuck.
The fit call determines whether five days can materially reduce the risk. If not, I will say so.