> ## 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.

# How Guhan scores prospects

> What "match" means when a Watchlist Agent finds someone.

Every prospect that comes through a Watchlist Agent gets a **match score** from 0 to 100.

## What the score means

* **80–100** — Strong match. All your filters check out — right title, right location, right industry, right company size.
* **50–79** — Partial match. Some filters match, some don't. Prospect lands but is flagged.
* **0–49** — Weak match. Filtered out. You can still see them under the "Filtered" toggle.

Prospects below the threshold (default: **50**) don't land in your List. They exist in your workspace as "filtered" but Guhan won't message them.

## What Guhan checks

<AccordionGroup>
  <Accordion title="Job title match">
    Compares the prospect's LinkedIn title against your list of target titles. Handles common variations — "VP of Sales" matches "Vice President Sales", "SVP Sales", "Head of Revenue", etc.
  </Accordion>

  <Accordion title="Location match">
    Does the prospect's LinkedIn location fall within your target regions? Handles city/state/country granularity.
  </Accordion>

  <Accordion title="Industry match">
    Does the prospect's company industry match your target industries? Uses LinkedIn's own industry taxonomy.
  </Accordion>

  <Accordion title="Headcount match">
    Is the company within your headcount range (e.g., 50–500 employees)?
  </Accordion>

  <Accordion title="HQ country match">
    If you set an HQ country filter, is the company actually headquartered there?
  </Accordion>

  <Accordion title="Signal match">
    Did the triggering signal actually fire on this prospect / company? (E.g., "posted about attribution" — does their recent post actually mention attribution?)
  </Accordion>
</AccordionGroup>

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

Some filters are **required** (must-match) and some are **weighted** (nice-to-have).

By default:

* **Title** and **Location** are must-match — fail either and the prospect is rejected outright.
* Everything else contributes to the score but isn't a hard gate.

You can change which filters are must-match on the Watchlist Agent's WHO panel — look for the **Required** / **Nice-to-have** chip next to each rule.

## Why prospects get filtered

The Watchlist Agent's detail page has a **Filtered** accordion showing every prospect that didn't make it, with a plain-English reason:

* "Wrong job title" — their title didn't match any of yours
* "Wrong location" — they're in a country you didn't target
* "Wrong industry" — their company's industry didn't match
* "Competitor employee" — they work at one of your competitors (auto-block)
* "Do Not Contact" — they're on your DNC list
* "Already in your workspace" — you found them via a different signal already

This is useful for debugging. If Guhan keeps rejecting people you think should qualify, the Filtered accordion tells you which rule is being strict.

## Tightening or loosening the threshold

Every workspace has a global **score threshold**. Default is 50 (moderately strict). You can override it per Watchlist Agent.

* **Tighten to 70** — fewer prospects, higher quality
* **Loosen to 30** — more prospects, more noise

Play with this after your first 100 prospects land — you'll see the shape of what you're getting.

## Related

<CardGroup cols={2}>
  <Card title="Prospect pipeline" icon="filter" href="/agents/watchlist-agents/prospect-pipeline">
    The full flow from LinkedIn search to landed prospect.
  </Card>

  <Card title="Watchlist Agent" icon="binoculars" href="/agents/watchlist-agents/overview">
    Configure the filters that drive scoring.
  </Card>
</CardGroup>