Normalize Customer Answers.prompt

# 61 - Normalize Customer / Domain Expert Answers

## Purpose

Normalize customer/domain expert answers into structured supplemental evidence that can be safely used by later impact analysis and FD update steps.

Customer answers are supplemental evidence. They do not modify the original BRD and must not be silently merged into source-derived artifacts.

This task must not generate requirements, business rules, FD content, DD content, or code design.

## Operating rules

- Use only the source artifacts listed in this prompt.
- Treat customer/domain expert answers as supplemental evidence, not as original BRD evidence.
- Do not invent meaning beyond the customer answer.
- Do not over-apply a narrow answer to unrelated requirements, rules, diagrams, or FD sections.
- Preserve customer answer wording where it affects meaning.
- Preserve product-specific terms, command names, option names, field names, acronyms, and code-like identifiers exactly.
- Follow `output/21_glossary.md` and `output/11_translation_policy.md`.
- Respect strikethrough/deprecated evidence rules from the pipeline:
  - strikethrough-only source evidence must not be treated as active behavior unless the source explicitly says otherwise.
  - if a customer answer appears to reactivate or reject strikethrough/deprecated evidence, mark the answer as requiring impact analysis.
- Do not expose internal prompt or pipeline details in customer-facing wording.
- Internal IDs may be used in this inventory for traceability.
- Write outputs in English unless source values must remain unchanged.

## Tasks

### Precondition

Use the latest available customer/domain expert Q&A file:

- Use `output/60_CUSTOMER_QA.md` as the question baseline.
- Use `input/supplemental/customer_answers.md` as the customer/domain expert answer source.
- If `input/supplemental/customer_answers.md` is missing or empty, stop with `No-Go`.

Use the latest available review context only to understand why the questions were asked:

- Prefer `output/55_FD_PATCHED_REVIEW_REPORT.md` if available.
- Otherwise use `output/52_FD_REVISED_REVIEW_REPORT.md` if available.
- Otherwise use `output/42_FD_REVIEW_REPORT.md` if available.

### Instructions

For each customer/domain expert answer:

1. Match it to the related customer-facing QID from `output/60_CUSTOMER_QA.md` when possible.
2. Preserve the original answer text in the inventory.
3. Normalize the answer into concise technical English.
4. Classify the answer type and affected area.
5. Link it to related source-derived items when identifiable:
   - requirements from `output/30_requirement_inventory.md`
   - business rules from `output/31_business_rule_catalog.md`
   - open questions from `output/32_open_questions.md`
   - terms from `output/21_glossary.md`
   - image evidence from `output/20_image_analysis.md`
6. Mark confidence and remaining ambiguity.
7. Do not update any downstream artifact in this task.

## Inputs

### Primary inputs

- `output/60_CUSTOMER_QA.md`
- `input/supplemental/customer_answers.md`
- `output/32_open_questions.md`
- `output/30_requirement_inventory.md`
- `output/31_business_rule_catalog.md`
- `output/21_glossary.md`

### Supporting inputs, open only when needed

- `output/55_FD_PATCHED_REVIEW_REPORT.md`, if available
- `output/52_FD_REVISED_REVIEW_REPORT.md`, if available
- `output/42_FD_REVIEW_REPORT.md`, if available
- `output/20_image_analysis.md`
- `output/12_normalized_evidence.md`
- `output/11_translation_policy.md`
- `output/10_document_inventory.md`

## Outputs

### Output file to create or update

Create:

- `output/61_customer_answer_inventory.md`

## Required output structure

Write `output/61_customer_answer_inventory.md` using this structure: