Back/Product/Claude
BeginnerProduct

How to Use Claude for AI-Powered Feature Prioritization

Keep feature requests in one Claude chat that estimates build time and scores customer happiness and growth impact. The result is a living table Daniel can scan when he has a few hours to work on Commutely.

How to Use Claude for AI-Powered Feature Prioritization

Daniel opens his Commutely feature tracker, explains the time and impact rubric, and shows how he uses its ranked table to choose a feature that fits his available time.

Before you start

What you need

  • A named product or project
  • A Claude project or persistent chat
  • A backlog of feature ideas
  • A simple scoring rubric

What you’ll make

A living Markdown table that captures every feature idea with an effort estimate, customer happiness score, growth score, rationale, and status.

Tools used

Step by step

The workflow

Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.

3 steps

Step01

Set Up Your Product Manager Persona

Create a persistent Claude chat for one product and define the backlog schema. Daniel uses estimated back and forth build hours plus separate one to three scores for customer happiness and growth impact.

Example prompt
Use this chat as the feature idea tracker for [product]. For every idea, maintain a Markdown table with: feature, user problem, estimated back and forth build hours, customer happiness score from 1 to 3, growth impact score from 1 to 3, rationale, dependencies, and status. Treat scores as estimates and keep all prior ideas unless I mark one removed.

Name the chat something memorable like "[Your Project Name] Feature Idea and Tracker" so you can easily return to it.

Step02

Log New Feature Ideas

Add each request in the same chat with enough detail to identify the user and desired behavior. Claude should update the existing row when an idea evolves, rather than creating a near duplicate.

Example prompt
New feature idea: add a standalone home screen widget that shows [core information] for [target user] without opening the app. Add or update this idea in the tracker, estimate effort, score both impact dimensions, and explain the main dependency or uncertainty.
Step03

Review the Prioritized Roadmap

Use the table as a menu when you have time to build. Ask Claude for candidates that fit the available hours, compare the rationale, then mark the selected item in progress and keep shipped work in the table for history.

The table makes it easy to spot high-impact, low-effort features you can tackle when you have a limited amount of time.

What good looks like

  • New ideas are added without losing earlier entries.
  • Every row uses the same scoring scale and time unit.
  • The table can be sorted by available time or impact.
  • Completed features remain recorded instead of reappearing as new work.

Build your next product with ChatPRD

Turn an idea into a PRD, user stories, and a plan.

Try ChatPRD free

After the steps

Runbook notes

How to recover when the loop fails and where human judgment helps.

Recover

If it goes sideways

Claude changes the scoring method between ideas
Put the exact columns, scales, and time unit in the first message and ask Claude to restate them above every regenerated table.
Impact scores sound precise without evidence
Require a one sentence rationale and mark scores as estimates. Add user evidence or usage data when it exists.
Older features disappear as the chat grows
Ask Claude to export the current table to a Markdown file after meaningful updates and use that file as the next session source.

Start shipping
better products.

Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.

Free to start
No credit card
SOC 2 certified
Enterprise ready