I need an adaptive pipeline review. Don't just answer standard questions — analyze the data and tell me what questions I should be asking this week based on what the pipeline actually shows.

**Pipeline data:**
[Paste your CRM export with: deal name, stage, value, days in current stage, days since last activity, number of contacts engaged, segment/vertical, next scheduled action, close date]

**Historical benchmarks:**
- Average days per stage for won deals: [if you have this data]
- Average win rate by segment: [if available]
- Typical deal cycle length: [your average]

**Analyze and produce:**

1. **Pipeline Composition Snapshot** — Total value by stage, by segment, by deal size tier. Flag any concentration risks (more than 40% of pipeline value in one segment, one deal, or one stage).

2. **Velocity Analysis** — Which deals are moving faster or slower than historical averages? Flag deals sitting in a stage for more than 1.5x the average.

3. **Stall Detection** — Deals with no activity in 14+ days. For each: how long since last activity, what was the last activity, and what's the risk of losing momentum.

4. **Single-Threading Risk** — Deals with only one contact engaged. These are one ghosting away from dead.

5. **Forecast Confidence** — For each deal, rate forecast confidence (High / Medium / Low) based on: stage, velocity, activity recency, and contact breadth. Calculate a confidence-weighted forecast.

6. **Top 5 Actions for This Week** — Ranked by revenue impact. Not "follow up on Deal X" — specific actions: "Schedule a technical evaluation call for Deal X to break the Stage 3 stall" or "Add a second contact to Deal Y to reduce single-threading risk."

7. **This Week's Questions** — Based on what the data shows right now, what are the three questions I should be discussing in my pipeline review meeting that I might not have thought to ask?