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

# How do I forecast my pipeline?

> Ask Claude to read your open deals and their stages and build a best/likely/worst forecast from your real pipeline data — no manual spreadsheet.

Someone asks what you will close this quarter and you either lowball it defensively or give an optimistic number pulled from thin air. A real forecast means weighting every open deal by stage and probability — a spreadsheet exercise you never sit down and do. Connect Claude to Klipy and it reads your actual deals and builds the projection.

## Example prompt

Copy this into Claude once it is connected to Klipy:

```text theme={null}
Read my pipeline and give me a forecast for this quarter — best case,
likely, and worst case — based on each deal's value, confidence, and
stage.
```

## What happens

<Steps>
  <Step title="Claude resolves your pipeline">
    Claude calls `list_pipelines` to get your pipelines and their stages — the structure a weighted forecast is built on.
  </Step>

  <Step title="It reads the deals">
    Claude calls `list_crm_records` for the `deal` type in each pipeline, reading each deal's `value`, `confidence`, `stage`, and `projectedCloseDate`, and uses `stage` to tell open deals from closed ones.
  </Step>

  <Step title="It builds the projection">
    Claude weights the deals by stage and confidence and reasons out best, likely, and worst scenarios from the real numbers — then reports them back to you. This is read-only; nothing is written to your CRM.
  </Step>
</Steps>

The whole workflow reads data and reasons over it, so it consumes no tokens.

## Good to know

* **Forecast quality depends on populated deals.** `value`, `confidence`, and `projectedCloseDate` are real fields on every deal — but they are optional. If your deals do not carry a value or a confidence, the forecast has less to weight, and the result is only as good as what you have filled in. Keeping those fields current is what makes the projection trustworthy.
* **Deals are scoped to a pipeline.** Listing deals requires a pipeline id, so Claude works pipeline by pipeline. If you run several, ask it to cover all of them or name the one you mean.
* **The scenarios are Claude's reasoning, not a stored model.** Klipy exposes the deal data; the best/likely/worst split is Claude weighting it. Two runs can phrase the split differently — treat it as a grounded estimate, not a fixed system number.

## Related

<CardGroup cols={2}>
  <Card title="Cold deal outreach" icon="flame" href="/docs/guides/mcp/use-cases/cold-deal-outreach">
    Act on the stalled deals the forecast surfaces.
  </Card>

  <Card title="Ask your data" icon="message-circle-question" href="/docs/guides/mcp/use-cases/ask-your-data">
    Ask follow-up questions about specific deals in the pipeline.
  </Card>

  <Card title="Tool reference" icon="wrench" href="/docs/guides/mcp/tools">
    Every tool Claude can call, grouped by scope.
  </Card>
</CardGroup>