Search Quality

Small changes to how you call search_humans can significantly improve result accuracy. These tips are based on patterns observed across thousands of searches.

Always provide company context

The single most impactful thing you can do. A name alone is ambiguous — company_name or company_domain narrows the search to the right person.

1
// Low accuracy — "David Solomon" matches many people
2
await search_humans({ name: "David Solomon" });
3
4
// High accuracy — company disambiguates
5
await search_humans({ name: "David Solomon", company_domain: "goldmansachs.com" });

Prefer company_domain over company_name

Domains are unique identifiers. Company names aren’t. company_name: "Front" is ambiguous. company_domain: "front.com" is not.

Use company_domain whenever you have it — for example, by extracting it from the lead’s email address.

Use location for multi-office companies

If the company has offices in multiple cities, add location to find the right person at the right office.

1
await search_humans({
2
  name: "Sarah Chen",
3
  company_name: "Google",
4
  location: "New York"
5
});

Understand the provide_email tradeoff

Setting provide_email: true adds 2-5 seconds of latency per search. It triggers domain discovery, email pattern generation, and SMTP verification.

• Use it when you need the email for send_meeting_invite or outreach
• Skip it when you only need profiles and location (e.g. for search_interests)

Split-engine strategy helps at scale

HumanConnection sends each query to a single search engine (Google or Brave), not all engines at once. This means:

• Lower rate limiting risk — no single engine sees all your queries
• More reliable results at high volume
• Occasionally, one engine may have better results for a specific query

If a specific search returns empty, retrying may route to a different engine.

Check confidence before proceeding

Don’t build a full meeting pipeline on a 0.2 confidence match. Add company_domain or location and search again.