Skip to main content

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 takes the guesswork out of TAM expansion. Instead of manually defining search criteria, you point Signal Ark at a handful of your best-performing, closed-won accounts and let the engine figure out what makes them tick — then go find more of them. The result is a ranked list of companies that mirror your ideal customers, sourced from both your existing database and the live web.

​
How it works

1

Select seed accounts

Choose between 3 and 10 accounts from your existing list to use as seeds. Pick accounts that represent your ideal customer — closed-won deals, high NPS scores, or strong product-market fit tend to produce the best results.
2

Profile generation

Signal Ark uses an LLM combined with People Data Labs (PDL) to extract the core DNA of your seed accounts. This includes their industry, tech stack, target audience, pricing model, and go-to-market motion — building a composite profile of what your best customer looks like.
3

Web discovery via Exa AI

The engine sends the composite profile to Exa AI for semantic search across the live web. Exa surfaces companies exhibiting the same characteristics as your seeds, not just companies that share the same keywords.
4

Scoring and filtering

Every candidate result is scored against your configured ICP criteria. Companies that don’t meet the minimum ICP fit threshold are filtered out before you ever see them.

​
Credit cost and caching

Running a Lookalike Discovery pipeline consumes 5 credits per run. To avoid charging you for redundant work, Signal Ark caches results for 24 hours. If you run the same set of seed accounts again within that window, you receive the cached results at no additional credit cost.
If you want fresh results after tuning your ICP scoring criteria, wait for the 24-hour cache to expire or change your seed account selection.

​
Reading your results

Results are split into two categories:

Internal results

Companies that already exist in your Signal Ark database but haven’t made it onto your active account list. These are your lowest-effort wins — one click adds them to your pipeline.

External results

Net-new companies discovered on the live web that have no record in your workspace yet. Promoting one creates a new account and triggers the standard enrichment and signal hydration flow.
Results within each category are ranked by similarity score — a composite measure of how closely each candidate matches your seed account profile across firmographics, tech stack, and go-to-market signals.

​
Adding results to your account list

To add a single company, click Promote to Account on any result card. For multiple results, use the checkboxes to select a group and click Bulk Promote. Before any credits are consumed for enrichment, Signal Ark shows you a preview with:
  • Total accounts to be added
  • How many already exist in your workspace (no enrichment cost)
  • How many are net-new (enrichment credit cost per account)
Confirm the preview to kick off enrichment and start monitoring.
External results that are net-new to your workspace will consume 1 credit per account for standard enrichment when promoted. This is separate from the 5-credit cost of running the Lookalike Discovery scan itself.