A useful weekly product review with AI starts with the current goal, saved decisions, evidence and unfinished work. Ask for one feasible next action and its tradeoffs. Keep “a document was written” separate from “the customer problem was solved.”

This guide is for a solo SaaS founder choosing how to spend the next week. You can use the worksheet manually or ask 1PC's Product Owner to read your saved product state. The example below is fictional.

Start with the product state you can support

Begin with what changed since your last review. A new feature draft, an accepted customer segment and a report of a user getting stuck are different kinds of information. Each should affect the next recommendation in a different way.

  • Goal: the customer or business outcome you are trying to improve now.
  • Accepted direction: the audience, scope and constraints you actually chose.
  • Evidence: observations, dates, sources and what they cannot establish.
  • Progress: completed checks, active work, blockers and items needing review.
  • Capacity: time and budget available after support and other commitments.

If this record does not exist, start with the founder context template. If you already have a working product, collect its current website and product documentation before asking for a fresh strategy.

Read the latest accepted choice rather than treating the longest document as authoritative. The product decision log helps distinguish an accepted direction from an attractive proposal.

Run a review that leads to a decision

Use the following worksheet. Fill unknowns explicitly; AI can organize missing information without supplying imaginary evidence.

# Weekly product review

Product and review date:
Current goal:
Accepted customer and scope:
Available hours and budget this week:

What changed:
- New observation or deliverable:
- Source and date:
- What it supports / what remains unknown:

Work state:
- Completed, with actual checks:
- In progress:
- Blocked or needs review:

Candidate next actions:
- Outcome, evidence, effort, dependency and uncertainty:

Choose one action:
- Why it matters now:
- What waits and why:
- Done criterion:
- What would change this choice:

Keep the review small enough to change your next action. Rewriting every strategy file each week makes it harder to see the actual change. Ask the assistant to point out conflicts and reuse current work before recommending a rerun.

Compare actions by the bottleneck they address

Suppose the fictional Follow-up Board helps freelance designers track unanswered proposals. You have ten hours this week and a goal of helping pilot users reach their first useful result. Your working notes contain a reported reminder issue, uncertain audience messaging and a draft for agency features.

Planning assumptions for a fictional product, not measured results
CandidateEvidence and uncertaintyBounded next action
Investigate missed remindersA supplied user report describes a missed reminder; the cause is unverified.Try to reproduce the issue and record the result before choosing a fix.
Clarify the landing-page promiseThe page emphasizes agencies while the accepted customer is a freelance designer.Draft a focused message correction using the current customer decision.
Build agency dashboardsA brainstorming document exists, but no accepted scope or customer evidence supports it.Keep it outside this week's delivery scope and record the open assumption.

A sensible first action could be investigating the reminder issue: it may block the customer outcome the product already promises. That is a reasoned choice under these assumptions, not an automatic rule that bugs always outrank discovery or marketing.

Compare the likely outcome, evidence strength, effort and dependencies in plain language. If you use a score, show the assumptions behind it. A precise number should not disguise uncertainty. Reserve time for review and customer follow-up before committing the whole ten-hour budget.

When the next action is a new build, use the MVP scope guide to define the smallest useful result. When it is a release message, use the feature-launch marketing plan to connect it to the accepted customer problem.

Ask Product Owner to read state before work

In 1PC, Product Owner is the entry point for each workflow. It reads relevant saved knowledge, decisions and progress, explains the current state, then selects guidance for the requested task. If a rerun or change has unclear consequences, it resolves that scope using your existing authorization.

For a status review, use a request like this:

@1pc Review my product using its saved knowledge, decisions
and progress. My goal is first-user value, with ten hours this week.

Tell me what is supported, what is unfinished and what is uncertain.
Recommend one next action with its tradeoffs and done criterion.
Explain which saved records informed it. Keep this review read-only.

Every User can load saved state and the available skills, with accepted private MCP calls charged to their Token Wallet. A review request does not establish implementation or a customer outcome.

After choosing the next task, request its specific deliverable and saving scope. Product Owner can use a built-in skill or an available published private skill. If you repeat the same review method, the custom product skills tutorial explains how to turn it into your own versioned checklist.

Keep the next conversation aligned with the decision

Ask explicitly to save a durable review and chosen next actions. Recommendations remain proposed until accepted. Preserve earlier decisions and their rationale when direction changes.

End the task with an observable checkpoint: what was checked, what was saved, what remains open and the next useful action. In a new connected conversation, ask Product Owner to read that saved state before continuing. The context continuity guide explains how to verify this handoff.

Repeat the review when new evidence or capacity changes the choice. The value is a current decision you can act on, with a record that helps the next task begin in the right place. For the broader sequence, return to managing product work in ChatGPT.

Choose the next useful product action.

Ask Product Owner for a read-only review. Every User can save authorized reports and continue work using Token Wallet credit.

Set up your product review ↗