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

# Lookalike Discovery

> Expand your Total Addressable Market (TAM) by finding clones of your best customers.

The **Lookalike Discovery** engine takes your highest-performing, closed-won accounts and scours the internet to find similar, high-potential companies that fit exactly the same profile.

## How It Works

Instead of manually guessing which industries or keywords to target, Lookalike Discovery automates the profiling and search process.

<Steps>
  <Step title="Select Seed Accounts">
    Choose 3 to 10 of your best, highest-intent accounts as the "seeds".
  </Step>

  <Step title="Profile Generation (LLM + PDL)">
    SignalArk extracts the core DNA of these companies—analyzing their tech stack, target audience, pricing model, and go-to-market motion.
  </Step>

  <Step title="Web Retrieval (Exa AI)">
    Using advanced semantic search via Exa AI, the engine scans the live web for companies exhibiting the exact same characteristics.
  </Step>

  <Step title="Scoring & Filtering">
    Results are filtered against your existing ICP and scored for relevance.
  </Step>
</Steps>

## Credit Usage & Caching

Running a full Lookalike Discovery pipeline is an intensive operation that consumes **5 credits** per run.

To optimize performance and credit usage, Lookalike results are cached for **24 hours**. If you run the exact same seed accounts within that window, you will receive the cached results without consuming additional credits.

## Interpreting Results

The results are split into two actionable categories:

1. **Internal Results:** Lookalike companies that *already exist* in your SignalArk database but were previously overlooked.
2. **External Results:** Net-new companies discovered on the web that you can instantly promote to your target account list with one click.