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

# Configure your ICP

> Get your ideal customer filters right.

**ICP** = Ideal Customer Profile. It's how you describe who you sell to. Guhan uses this to search LinkedIn + to score every prospect it finds.

## The four ICP fields

<CardGroup cols={2}>
  <Card title="Job titles" icon="user-tie">
    Their role. What do they do at the company?
  </Card>

  <Card title="Location" icon="location-dot">
    Where they live or work.
  </Card>

  <Card title="Industry" icon="building">
    What industry their company is in.
  </Card>

  <Card title="Company size" icon="users">
    Headcount range.
  </Card>
</CardGroup>

Plus optional filters:

* **HQ country** — stricter than person-location; the *company* has to be headquartered there
* **Funding stage** — company's most recent funding round
* **Custom rules** — via the Branch node in sequences (Advanced)

## Getting job titles right

The most impactful filter. Start broad, narrow later.

**Broad approach (recommended for first Watchlist Agent):**

```
VP of Sales
Head of Sales
Chief Revenue Officer
CRO
```

Guhan will auto-expand each into variants (Vice President Sales, SVP Sales, Head of Revenue, etc.).

**Narrow approach (for tightly-targeted campaigns):**

```
VP of Sales Operations
Head of Sales Enablement
Director of RevOps
```

More specific = fewer prospects but higher precision.

<Tip>
  Auto-variants are shown as ✨ chips under each rule row. Delete any that don't fit your ICP.
</Tip>

## Location tips

<AccordionGroup>
  <Accordion title="Country-level">
    "United States" catches every US resident. Good for broad geo targeting.
  </Accordion>

  <Accordion title="State/region-level">
    "California" or "Bay Area" catches everyone in that state/region. Better for pipeline focused on specific tech hubs.
  </Accordion>

  <Accordion title="City-level">
    "San Francisco" catches SF-metro only. Narrowest.
  </Accordion>

  <Accordion title="Multiple locations">
    Add them as separate chips. Logic is OR — a prospect matches if they're in any of the locations.
  </Accordion>
</AccordionGroup>

## Industry tips

LinkedIn's industry taxonomy is specific. Common ones:

* **Software Development** — the "SaaS" bucket
* **Financial Services** — fintech, banks, brokers
* **Marketing Services** — agencies + consultancies
* **Information Technology & Services** — IT services + integrators
* **Computer & Network Security** — cybersecurity

Full list is searchable in the picker. Multi-select with OR logic.

<Warning>
  **"SaaS" isn't in LinkedIn's taxonomy.** Use Software Development + narrow via signals or company size.
</Warning>

## Company size tips

Guhan reads LinkedIn's employee count for each company:

| Range        | Typical company shape    |
| ------------ | ------------------------ |
| 1–10         | Solo founders, pre-seed  |
| 11–50        | Seed / Series A          |
| 51–200       | Series B / early scaleup |
| 201–500      | Mid-market               |
| 501–1,000    | Larger mid-market        |
| 1,001–10,000 | Enterprise               |
| 10,001+      | Fortune 500              |

Leave both ends blank → no size filter.

## Must-match vs. nice-to-have

You can mark each filter as **Required** or **Nice-to-have**:

* **Required** — prospects failing this filter are rejected
* **Nice-to-have** — contributes to the match score but doesn't reject

By default, **Titles** and **Location** are Required (they're the definitional filters). Everything else is Nice-to-have.

Change with the chip next to each rule row on the Watchlist Agent WHO panel.

## Score threshold

Every prospect gets a match score 0–100. Below the workspace threshold (default 50) = filtered out.

You can override the threshold per Watchlist Agent:

* **Tighter (70+)** — fewer prospects, higher quality
* **Looser (30+)** — more prospects, wider funnel

Find it on the Watchlist Agent detail page → Settings.

## Debugging your ICP

If your first sweep found zero prospects:

1. **Titles too narrow?** Try broader — "Sales Leader" instead of "VP of Enterprise Sales"
2. **Signals too tight?** Remove or loosen your signal
3. **Industry mismatched?** Pick a broader LinkedIn category
4. **Company size mismatched?** Widen the range

If your first sweep found 200+ prospects of dubious quality:

1. **Narrow titles** — remove overly broad chips
2. **Add signals** — turn on funding OR hiring to focus on active buying windows
3. **Tighten the score threshold**

## Related

<CardGroup cols={2}>
  <Card title="ICP scoring" icon="chart-simple" href="/agents/watchlist-agents/icp-scoring">
    How the score is calculated.
  </Card>

  <Card title="Signal catalog" icon="bolt" href="/agents/watchlist-agents/signal-catalog">
    Every signal with examples.
  </Card>
</CardGroup>