# AI FAQ Generator

## Objective

Produce short question and answer pairs from a document you have already approved.

A set of concise question-and-answer pairs generated from one nominated source document, each traceable back to the passage it came from.

## Before You Begin

This feature is being added to an application that already exists and already
works. Do not scaffold a new project, and do not assume a blank slate.

Inspect the codebase first and establish:

- The existing application structure and where code of this kind already lives.
- The framework and version in use.
- The existing design system — colours, spacing, typography, and component conventions.
- Existing UI components you can reuse instead of writing new ones.
- The existing database structure, if this feature needs to persist anything.
- The existing authentication and authorization system, if this feature is user-scoped.
- Dependencies already installed, so you don't add a library that duplicates one.
- The existing test setup and conventions.

Only start writing code once you understand the above. If the application
already implements part of this feature, extend it rather than replacing it.

## Implementation Instructions

1. Require the user to nominate a single approved source — a help article, a product page, a policy document — and generate only from that. Do not blend several documents or fall back to general knowledge when the source is thin.
2. Store each generated pair with a reference to the passage in the source that supports it, so a reviewer can confirm the answer without rereading the whole document.
3. Keep answers to a few sentences and link to the fuller section of the source for anything longer. An FAQ that restates the entire document is just a worse copy of it.
4. Compare new pairs against the existing set before saving, and merge or discard questions that ask the same thing in different words rather than accumulating near-duplicates.
5. Record the version of the source each pair was generated from, and mark pairs as out of date when the source changes rather than silently rewriting text a person already approved.

## UI and UX Requirements

Match the application's existing design system exactly. Reuse its components,
spacing, and typography. This feature should look like it was always there.

## Responsive Requirements

Works on mobile, tablet, and desktop. Touch targets are large enough to hit on a
phone, and nothing overflows horizontally at 320px.

## Accessibility Requirements

- Fully keyboard navigable.
- Correct semantic elements and ARIA roles.
- Visible focus states.
- Meets WCAG AA contrast.
- Dynamic changes are announced to screen readers.
- Respects prefers-reduced-motion.

## Edge Cases

- Any claim not present in the source is a fabrication, however reasonable it sounds. Drop pairs that cannot be traced to a passage instead of publishing them with a caveat.
- Generating from a document without generating for an audience produces questions nobody asks. Constrain the set to what a customer would plausibly type into search, and let the reviewer delete the rest easily.
- Long answers defeat the format. Enforce a length limit in the output and link out for detail rather than truncating mid-sentence.
- Overlapping questions clutter the list and split search results. Deduplicate on meaning, not on exact string match.
- When the source is edited, existing pairs must be flagged as stale and offered for regeneration. Leaving them unmarked means the FAQ quietly contradicts the document it came from.
- A short, structural, or mostly visual source yields nothing useful. Return an empty result with an explanation rather than padding the list.
- Refusals, timeouts, and malformed structured output must leave the existing approved pairs untouched.
- This feature generates short pairs from an approved document; producing full articles from solved support issues belongs to AI Knowledge Base Drafting, and the two must write to different content types.

## Testing

Exercise the feature end to end in the running application. Cover every edge case
above, then run the existing test suite and confirm nothing regressed.

## Acceptance Criteria

- [ ] Every pair traces to a passage in the single nominated source.
- [ ] Claims absent from the source are dropped rather than published.
- [ ] Answers respect a length limit and link to fuller guidance in the source.
- [ ] Near-duplicate questions are merged or discarded before saving.
- [ ] Pairs are versioned against the source and marked stale when it changes.
- [ ] An unsuitable source returns an empty result with an explanation.
- [ ] Failed generation leaves previously approved pairs unchanged.
- [ ] The feature matches the existing design system.
- [ ] No existing functionality is broken.

## Adaptation Rules

- Match the existing design system. Do not introduce a new colour palette,
  spacing scale, or component library.
- Reuse existing components and utilities wherever they fit.
- Follow the naming, file layout, and code style already present.
- Do not upgrade, replace, or remove existing dependencies to make this
  feature fit. Adapt the feature to the app, not the app to the feature.
- Do not break existing functionality. If a change is genuinely required in
  existing code, make the smallest one that works and say so.
- If something in these instructions conflicts with how the application is
  built, follow the application and explain the deviation.

## Final Verification

Before you report the work as done:

1. Re-read the acceptance criteria above and check each one against what you
   actually built.
2. Run the application and exercise the feature end to end.
3. Run the existing test suite and confirm you have broken nothing.
4. Check the feature on mobile, tablet, and desktop widths.
5. Check keyboard navigation and focus handling.
6. Summarize what changed: files added, files modified, and anything you
   deliberately did differently because of how this application is built.

If any acceptance criterion is unmet, fix it before reporting completion.