Available Context

@_platform-references/org-variables.md

AI Ark Semantic Search

Goal

Search for companies using natural language descriptions. AI Ark does not have a native semantic search endpoint — this skill converts natural language into keyword-based filters.

Credit Cost

Each search request costs ~2.5 credits (uses the company search endpoint with keyword filters). There is no free preview mode. Always warn the user before executing.

Required Capabilities

• AI Ark API: Keyword-based company search via ai-ark-semantic edge function

Inputs

• natural_language_query: Plain English description of target companies (required)
• max_results: Cap on results (default: 50)

Execution

1. Warn the user: "This search will cost ~2.5 AI Ark credits. Proceed?"
2. On confirmation, call ai-ark-semantic with the natural language query
3. The edge function extracts keywords and searches using account.keyword and account.productAndServices filters
4. Present results as a table for review
5. User can refine the query conversationally (each refinement costs another ~2.5 credits)

Output Contract

Return a table with columns:

• Company Name, Domain, Industry, Employee Count, Employee Range, Location, Description, Technologies

How It Works

The natural language query is split into keyword phrases (by commas, "and", semicolons) and searched against company keywords, descriptions, SEO tags, names, and industry fields. A query like "companies building AI tools for HR in Europe" becomes keyword filters for "companies building AI tools for HR" and "Europe".

Refinement Loop

The copilot should help users iterate:

1. Show results: "These are mostly enterprise companies"
2. User refines: "Focus on startups under 200 employees"
3. Re-search with refined query (costs another ~2.5 credits)
4. Additional filters can be passed via additional_filters for structured criteria