> ## Documentation Index
> Fetch the complete documentation index at: https://closedloop.sh/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# PM Prep

> Pre-call discovery brief for product managers: learning goals, knowledge gaps, adaptive pattern recognition, and questions from real customer transcripts.

Pre-call discovery brief designed for how product managers actually think. Not account health: learning goals, knowledge gaps, adaptive pattern recognition, and questions generated from real customer data.

## Usage

```
/closedloop-pm-prep Acme Corp
/closedloop-pm-prep acme.com
```

Type a customer name or domain. The skill assembles a discovery brief from feedback history, cross-customer patterns, conversation transcripts, and strategic context.

## What You Get

* **Learning goal**: what this specific call can teach you, framed as a hypothesis to test or a knowledge gap to fill. Written after reading all evidence.
* **What they're trying to do**: their underlying needs reframed as outcomes, not feature requests. Workarounds highlighted with sophistication level (spreadsheet hack vs. custom tool vs. hired a person).
* **What we know from others**: adaptive relative positioning showing where this customer's pain ranks against the full customer base. No fixed thresholds: rank and percentage tell the story.
* **What we don't know**: the highest-value section. Knowledge gaps this call could fill. Turns a status update into a research instrument.
* **Suggested questions**: data-grounded, Mom Test compliant. "Walk me through the last time..." not "Would you use...?" Structured as must-ask (2-3), should-ask (2-3), if-time (2).
* **Segment lens**: where this customer sits relative to peers, preventing over-indexing on one voice.

## How It Works

The skill launches 5 parallel research agents:

1. **Workflow + outcomes**: reads all feedback, reframes requests as underlying needs, surfaces workarounds
2. **Cross-customer patterns**: searches the same topics WITHOUT customer filter to find how many others share the pain, computes rank and percentage against total customer base
3. **Segment context**: finds peers in the same segment, compares concerns to prevent bias
4. **Conversations**: reads last 2 call transcripts for the customer's voice, open threads, and workflow context
5. **Strategic context**: competitor mentions decoded as underlying needs, satisfaction and churn indicators

## PM Prep vs. CSM Prep

| Dimension         | PM Prep                             | CSM Prep                    |
| ----------------- | ----------------------------------- | --------------------------- |
| Opens with        | Learning goal (what to discover)    | Headline (account judgment) |
| Shows             | Underlying needs, knowledge gaps    | Landmines, open threads     |
| Cross-customer    | Adaptive rank vs. full base         | Count of other customers    |
| Questions         | Mom Test, data-grounded             | Not included                |
| Workarounds       | Highlighted as opportunity evidence | Not surfaced                |
| Segment context   | Always shown                        | Not included                |
| Account health    | Not shown                           | Central                     |
| Deal/revenue data | Not shown                           | Central                     |

## Design Principles

**Outcomes, not features.** "Customer requested CSV export" becomes "Cannot incorporate data into existing reporting workflows." The PM sees the job, not the solution.

**Adaptive thresholds.** "114 of 1,997 customers (5.7%), #1 platform-wide" tells the PM this is a market signal. No magic numbers like "10+ = important."

**Knowledge gaps drive the call.** "What we don't know" is mandatory. It transforms a customer conversation from status update into research instrument.

**Questions from data, not templates.** Every suggested question traces to a specific insight, context record, or open thread. Never generic.

**Workarounds = gold.** If they built their own solution, the need is real. The sophistication of the workaround tells you severity.