Feature request triage and RICE scoring with stated assumptions
Groups a raw list of feature requests into underlying problems, de-duplicates them and scores each with RICE, showing every assumption so the team can argue about inputs instead of outcomes.
Prompt in Product by AISkillHub Library
What it does
Feature request lists grow faster than anyone can read them, and many entries are the same problem described in different words. This prompt groups requests by the underlying user problem, merges duplicates, and gives each problem a RICE score (Reach, Impact, Confidence, Effort) where every input is written down with its reasoning.
**Use it when** you have a backlog export, a spreadsheet of requests from sales and support, or a pile of feedback and need a defensible shortlist for planning.
**What you get** - Problem groups: each with a problem statement, the requests it covers, and the number of distinct requesters. - A RICE table with Reach, Impact (on a 0.25 / 0.5 / 1 / 2 / 3 scale), Confidence (50 / 80 / 100%), Effort (person-weeks) and the resulting score. - An assumptions column explaining each input, and which inputs need data you did not provide. - Requests that do not fit any group or are out of strategy, listed separately rather than silently dropped. - A suggested shortlist with a sentence on why each made it.
**What it will not do** - It will not estimate engineering effort with false precision. Unless you supply effort estimates, effort is given as a rough T-shirt range converted to person-weeks and marked as a placeholder for engineering to replace. - It will not invent reach numbers. If you do not supply usage or customer counts, it uses request counts and says so. - It will not decide strategy for you. If you provide strategic goals, it flags requests that conflict with them.
**Example** Requests like "export to Excel", "CSV download for reports" and "need data in our BI tool" are grouped as "Users cannot get report data out of the product for their own analysis", covering 3 requests from 11 customers.
How to use it
## How to use
1. Open `prompts/feature-request-triage.md` and copy everything below the title line. 2. Replace each `{{variable}}` with your own content. Paste long material (notes, transcripts, tables) directly in place of the variable; keep the surrounding tags so the model can tell inputs apart. 3. Start a new chat in ChatGPT, Claude or Gemini and paste the filled prompt as your first message. 4. Read the output against the quality checks at the end of the prompt before you use or send it.
## Variables
- `{{requests}}`: the raw requests, one per line, ideally with requester or account and source, e.g. `Acme (sales call): wants Excel export`. - `{{reach_data}}`: any usage or customer-count data to estimate reach, e.g. `2,300 active accounts, 400 use reports weekly`. Optional. - `{{effort_estimates}}`: known effort estimates from engineering, if any. Optional. - `{{strategy}}`: current goals or themes for the period, e.g. `reduce churn in mid-market`. Optional. - `{{time_period}}`: the period reach is measured over, e.g. `per quarter`.
If you do not have a value for an optional variable, write `none` rather than deleting it, so the prompt still reads correctly.
## Reusing it
To reuse it often, save the prompt (with the variables left in) as the instructions of a Claude Project, a custom GPT or a Gemini Gem, then paste only the inputs in each new chat.
Files
Save each file at the path shown, relative to your project root.