# Weekly Pipeline Health Check Playbook

This playbook provides a high-level strategic audit of your sales pipeline every week. It identifies stalled deals, tracks conversion velocity, and flags missing data, ensuring you spend your time on the opportunities most likely to close.

## 1. Prompt Pack

These prompts guide your AI in auditing your CRM or project pipeline to identify risks and opportunities.

### Core Pipeline Health Audit Prompt

```
"You are a Sales Operations Director. Review the current sales pipeline data for the week ending {Friday Date}. Your task is to generate a 'Pipeline Health Report' that includes:

- **Volume Metrics:** Total number of active deals and total pipeline value.
- **Velocity Check:** Identify any deals that have not changed stage in more than 14 days. List these as 'Stalled Deals'.
- **Risk Assessment:** Flag any deals where the 'Next Step' is missing or past due.
- **Conversion Forecast:** Based on current stages and historical win rates, what is the 'Weighted Forecast' for the next 30 days?
- **Action Recommendations:** Suggest 3 specific actions to improve pipeline health this week (e.g., 'Follow up on Stalled Deal X', 'Cleanup dead leads in Stage Y').

Save the report to memory/sales/pipeline-health/weekly-{YYYY-MM-DD}.md."
```

### Prompt Variations:

*   **The 'Leakage' Audit Prompt:**
    ```
    "Analyze the deals marked as 'Lost' in the last 30 days. Identify the top 3 reasons for loss (e.g., Price, Timing, Competitor). Recommend one change to our sales process to address the most common reason."
    ```
*   **The 'Big Fish' Alert Prompt:**
    ```
    "Identify the top 5 deals by value currently in the pipeline. For each, summarize the current status and suggest a 'High-Touch' action the founder or CEO should take this week to move it forward."
    ```
*   **The 'Data Integrity' Cleanup Prompt:**
    ```
    "Scan all active deals for missing fields (e.g., Close Date, Deal Source, Primary Contact). Generate a checklist of specific deals that need data cleanup to ensure accurate reporting."
    ```

## 2. Setup Guide

To implement the Weekly Pipeline Health Check Playbook:

1.  **Save the Playbook:** Save this content as `playbooks/weekly-pipeline-health-check.md`.
2.  **Create Cron Configuration:** Create `playbooks/pipeline-health-cron.json` (see Config Snippet below).
3.  **Integrate Pipeline Source:**
    *   **CRM Export:** If you use a CRM (HubSpot, Pipedrive, etc.), set up an automated weekly CSV export to a folder the agent can read.
    *   **Shared Sheet:** If using a spreadsheet, ensure the agent has access via the relevant integration or file-drop.
4.  **Activate Cron Job:** Once the JSON config is added, OpenClaw will run the audit every Friday afternoon.
5.  **Review the Report:** Use the 'Action Recommendations' to set your priorities for the following Monday.

## 3. Config Snippet

This JSON snippet defines a weekly cron job that triggers the pipeline audit.

```json
{
  "cron": [
    {
      "name": "Weekly Pipeline Health Audit",
      "schedule": "0 15 * * 5",
      "task": "Every Friday at 3:00 PM, scan the active deals in 'shared-context/sales/pipeline.json'. Use the 'Core Pipeline Health Audit Prompt' from the 'Weekly Pipeline Health Check Playbook' to generate the weekly report. Notify the COO once the report is saved to memory.",
      "agent": "cmo",
      "model": "gemini-flash"
    }
  ]
}
```

## 4. Permissions Checklist

Your agent needs:

*   **Read Access:**
    *   Pipeline data file or CRM API.
    *   Historical win rate data (for weighted forecasting).
*   **Write Access:**
    *   `memory/sales/pipeline-health/` (to store weekly reports).
*   **Execution Permissions:**
    *   Ability to perform basic calculations (totals, averages, weighted values).

## Safety Check (Required)

- Ensure pipeline data is stored securely and not exposed to unauthorized sessions.
- Verify that financial forecasts are used as a guide, not as guaranteed revenue projections.
- Periodically check the 'Data Integrity' cleanup prompt to ensure the AI has the quality information it needs to be accurate.