As of 13 August 2026, AI can prioritise product features.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open your product backlog, roadmap or feature list and copy each proposed feature into a separate line.
   2. Gather current customer feedback, support themes, sales notes, product usage data and research, and label each item with its source and date.
   3. Write down your current business goals, target customers, delivery constraints, dependencies and any commitments that affect timing.
   4. Paste the context, feature list and evidence into the prompt, replacing every bracketed slot with your own information.
   5. Ask the model to produce the scoring table, both rankings, the calculations, the assumptions and the missing-information questions specified in the prompt.
   6. Compare every evidence claim and numerical input in the output against your source documents and correct anything that is missing, misread or invented.
   7. Ask a product or commercial colleague to challenge the scoring criteria and assumptions, then agree the ranking and next steps in your roadmap or planning document.

   Prompt

   Act as a product strategy analyst for a UK business. Help me prioritise the feature ideas below without inventing evidence, customer needs, costs, revenue, deadlines or technical facts.
   
   Business context:
   - Business and product: [describe them]
   - Target customers: [describe them]
   - Current business goals: [list them]
   - Constraints, such as team capacity, compliance, dependencies or deadlines: [list them]
   
   Feature ideas:
   [paste one feature per line, with any known evidence, estimated effort, dependencies and risks]
   
   Evidence:
   [paste customer feedback, support themes, usage data, sales notes, research or other sources, labelling each source]
   
   Create a table with one row per feature and these columns: customer problem addressed, evidence strength, expected business value, reach or affected customers, strategic fit, urgency, delivery effort, dependencies, risks, confidence, and missing information. Use a simple scoring scale from 1 to 5 for value, reach, strategic fit and urgency, and score effort from 1 to 5 where 5 is hardest. Do not treat estimates as facts.
   
   Give two rankings: one based on business value relative to effort, and one based on strategic urgency. Explain the scoring formula in plain English, show the calculation for every feature, identify where the ranking changes because of assumptions, and propose the smallest useful next step for the top three features. Include a list of questions I should answer before committing resources. Separate evidence from your judgement. If the evidence is insufficient, say so and give a provisional ranking rather than filling gaps with guesses. End with a short decision brief for a product meeting, but state clearly that the final prioritisation remains with the business.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Hand it to a person

   The distant third

   A person who owns the outcome does this end to end, worth it when the failure is dear.