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

# Signals & Signal Agents

> 24/7 LinkedIn engagement monitoring — auto-add in-market buyers to your pipeline

A **Signal Agent** runs 24/7, watches LinkedIn posts that match your sources (keywords, influencers, competitors, your own content), and surfaces the people who comment or react. Every engager is scored against your ICP — only matches land in your list.

The closest Komo gets to inbound: you don't pay for the prospect, and you don't have to find them.

## Two tabs at `/signals`

* **Agents** — your Signal Agents with status, leads found, next-run time
* **Feed** — every match: triggering post, engager profile, ICP score. Filter by agent, source, label.

Pause, resume, or run any agent on demand.

## How an agent works

1. **Discover posts** from each signal source
2. **Collect engagers** (comments + optional reactions)
3. **Score against ICP** (titles, locations, industries + free-text criteria)
4. **Add to list** — matched engagers become contacts, with the triggering post attached

Runs **twice a day** by default. Posts dedupe within 24h.

## Creating an agent

**Lists → New list → Signal Monitoring** (or **Signals → New agent**). Four parts:

### 1. Signal sources (pick any combination)

| Source                    | Provide                         | When to use                           |
| ------------------------- | ------------------------------- | ------------------------------------- |
| **Engagement keywords**   | Tags (`cold email`, `ISO27001`) | Widest net                            |
| **Influencer profiles**   | Profile URLs                    | Audience of voices your buyers follow |
| **Competitor companies**  | Company page URLs               | Anyone engaging is warm for you       |
| **Your LinkedIn profile** | Your profile URL                | Re-engage your own audience           |
| **Your company page**     | Your company page URL           | Engagement with company posts         |

### 2. ICP filters

* **Job titles** (e.g. `VP Sales`, `RevOps Lead`)
* **Locations** (e.g. `United States`, `EMEA`)
* **Industries** (e.g. `SaaS`, `Fintech`)
* **Matching mode**:
  * **High precision** — fewer, better leads (default)
  * **Discovery** — broader net

### 3. Caps & cost controls

| Cap                  | Default | Limits                                  |
| -------------------- | ------- | --------------------------------------- |
| Posts per source     | 10      | Recent posts scanned per source         |
| Raw engager cap      | 400     | Unique engagers scraped per run         |
| Apify enrichment cap | 200     | Profiles enriched per run               |
| Daily lead cap       | 100     | Max engagers added per day              |
| Fetch reactions      | on      | Reactions + comments (\~2× scrape cost) |

### 4. Advanced (optional)

* **Additional AI criteria** — free-text rules for the scorer (`"Prefer founders. Skip students. Prioritize anyone hiring SDRs."`)
* **Exclude service providers** — drops consultants, agencies
* **Exclude Open-to-Work** — drops profiles flagged on LinkedIn

A **Run immediately** checkbox kicks off the first run as soon as you save.

## Feed view

Each match card shows: triggering post, engager profile, engagement type, signal source, ICP score and reasoning. Click an engager to add to a different list, push to inbox routing, or mark as read.

## Example: high-quality outbound for a SaaS company

```
Name: "VP Sales engaging with cold-email content"

Sources:
  Engagement keywords:  "cold email", "outbound sequence", "SDR playbook"
  Influencer profiles:  <2 sales thought leaders>
  Competitor companies: <2 competitors>

ICP filters:
  Titles:      VP Sales, Head of Sales, CRO
  Locations:   US, Canada
  Industries:  SaaS, IT
  Mode:        High precision

Caps:
  Daily lead cap: 100

Advanced:
  Additional: "Prioritize anyone mentioning hiring, building, or measuring an outbound motion."
  Exclude service providers: ON
```

Expected output: \~20–60 leads/day, every one publicly in-market.

## Pause, run, delete

From the **Agents** tab menu: **Run now**, **Pause / resume**, **Delete**.

## Connected to

[Contacts](/product/contacts) (matched engagers) · [Accounts](/product/accounts) (auto-created parent) · [Lists](/product/lists) (the bound list) · [Campaigns](/product/campaigns) (attach list → auto-enroll new matches) · [Inbox](/product/inbox) (replies)

## Tips

<Tip>
  **Start with high precision.** Discovery mode is noisier — only loosen if daily lead count is too low.
</Tip>

<Tip>
  **Engagement keywords cast the widest net.** Influencer/competitor sources are narrower but higher-quality.
</Tip>

<Tip>
  **Use Additional criteria aggressively.** Structured filters can't capture nuance like "anyone complaining about Outreach."
</Tip>

<Tip>
  **Leave Fetch reactions on.** A typical post has 5–10× more reactions than comments — worth the 2× scrape cost.
</Tip>