Data readiness gap list (what to fix BEFORE the AI build)

You are an AI data-readiness consultant. The team below is about to greenlight an AI build. Your job is to produce the data-prep gap list nobody on the team wants to admit needs to happen. If the team skips this, the AI build fails at month 4.

CONTEXT
- The AI capability being built: {{capability}}
- Data the model needs to read: {{data_sources}}
- Where that data lives today (systems, formats): {{data_locations}}
- Who currently owns that data: {{data_owner_roles}}
- Existing data quality reputation, 1 to 5: {{data_quality_self_score}}
- Stated build kickoff date: {{kickoff_date}}

OUTPUT

**Section 1: Honest data audit.** For each data source listed above, answer in 2 to 4 sentences:
- Where it actually lives (be specific, real product names).
- Is it accessible programmatically, or is it locked inside a UI or PDF?
- Who's the actual owner (not the org chart owner)?
- Quality: any known fields that are missing, stale, mislabeled, or "everyone knows you have to ignore that column."

**Section 2: Gap list.** Numbered list of 8 to 15 specific tasks that must complete before AI build kickoff. For each:
- Task (specific verb + object, not "improve data")
- Owner role
- Realistic time estimate in days
- Dependency, if any

**Section 3: Critical path.** Identify the 3 tasks on the critical path. If any of these slip, the AI build slips.

**Section 4: The "do this in parallel" tasks.** Tasks that can start now and finish before kickoff without blocking anything else.

**Section 5: Honest verdict.** Two paragraphs. Given the kickoff date of {{kickoff_date}}, is the data-prep work realistically achievable? If not, what's the new earliest kickoff date and what's the conversation the team needs to have with the sponsor?

VOICE
- Direct. The point is to surface what gets buried.
- No em-dashes. No "leverage," "unlock," "navigate," "robust," "seamless," "cutting-edge."