Prioritise product requests with AI
Compare product requests using evidence-labelled scores and a separate strategic view before choosing the next cycle’s work.
Prioritise product requests with AI by making the evidence behind each ask visible. The pm-portfolio-v1 skill preserves the original wording, separates independently decidable requests and compares product value with a distinct strategic view.
You receive a request table, a batch verdict and a proposed sprint or quarterly direction. A low numerical score does not automatically eliminate a strategic opportunity. The two views remain separate for the owner’s decision, and the skill does not start engineering work.
What it does
- Break a mixed batch into independently decidable requests
- Connect each ask to the product’s main outcome metric
- Separate verified request evidence from assumed benefits
- Calculate RICE with visible inputs and evidence confidence
- Expose disagreements between product and strategic value
- Propose the next cycle’s work with reasons for exclusions
Install and use
- Load the full Product2Prod package in a file/shell-capable agent with Python. Provide the request batch, main metric, boundaries and writable output location.
- Ask: “Compare these twelve requests for next quarter. Use fewer failed bookings as the goal and keep evidence gaps explicit.”
- Inspect the request table, claim sources and scoring inputs. Resolve missing goals or evidence rather than accepting a ranking built on guesses.
- At the human gate, approve, revise or block the recommendation. Collect the verdict files; spreadsheet output additionally requires a defined workbook form in the project pack.
FAQ
Source
Made by Nader Elewa. Original skill: pm-portfolio-v1, published on GitHub under the MIT licence.
View on GitHub