ai-feature-prd

1# AI Feature PRD Skill
2 
3AI features break the normal PRD because the system is probabilistic: it will be wrong sometimes, and
4the product must be designed around that, not in denial of it. This skill extends a standard PRD with
5the AI-specific sections that decide whether the feature is trustworthy — the UX of uncertainty, the
6eval bar, guardrails, and what happens when the model is wrong.
7 
8## Required Inputs
9 
10Ask for these only if they aren't already provided:
11 
12- **The user problem** and why an AI/probabilistic approach fits it (vs. deterministic rules).
13- **What "good" looks like** to the user, and the cost of a wrong answer (low-stakes vs. high-stakes).
14- **Inputs available** — context/data the model can use; privacy constraints.
15- **Trust level needed** — can the user verify the output, or must it be near-perfect?
16 
17## Reads from / Writes to the Brain
18 
19If a [`professional-brain`](../professional-brain/SKILL.md) exists, read `context.md` (product, users, voice)
20and `knowledge/strategy.md` first; write the feature to `entities/` and any scoping decision to `decisions/`,
21each provenance-tagged.
22 
23## Output Format
24 
25### AI Feature PRD: [feature]
26 
27**1. Problem & why AI** — the user problem, and why a model (not rules) is the right tool. If rules would do, say so.
28 
29**2. Experience** — the core flow, and crucially the **UX of uncertainty**: how confidence is shown, how the user verifies/edits, and how errors are made cheap to recover from. AI features live or die here.
30…