Build a verified B2B lead list with your AI agent
Five candidates, two checked roles and two independently verified work emails. Follow the calls and the $0.07980 data bill.
Sales and GTM ·
Start with a company and a role
A useful prospecting list needs more than a name and an email-shaped string. It should explain who the person works for, what they do now, where that information came from and what an email verifier actually checked.
For this example, we asked for marketing and growth people at Linear, using linear.app as the company filter. The search returned five candidates. We selected two product marketing contacts for a closer check, read their LinkedIn profiles, found their work emails and ran a separate verification call for each address. This is a sample at one named company; it does not measure coverage across an entire market.
| Contact | Current role | Email check |
|---|---|---|
| Emily Brody | Product Marketing at Linear | Valid; accept_all: false |
| Brian Hendery | Product Marketing at Linear | Valid; accept_all: false |
Keep finding and verifying separate
The people search used anyapi.people_search.quickenrich with the employer’s domain. We then used anyapi.linkedin.profile to check the two selected candidates’ current positions. A search title can be a useful lead, but a current profile is a separate piece of evidence. Neither proves that someone owns a particular budget.
Hunter’s email finder returned both addresses with inferred provenance and accept_all: true. A domain that accepts every address can make a guessed mailbox look convincing. We kept those finder flags and passed each returned address to hunter.email_verifier. Both dedicated checks returned valid, with smtp_check: true and accept_all: false. That later result is the basis for the verification column. It is a point-in-time deliverability check, not a guarantee of a reply.
Keep unknown, catch-all and invalid results in separate columns or a separate list. Do not silently promote them into the verified count. In the public example file, the addresses are masked; the private working results retain the full outputs.
What the seven calls cost
The search and two role checks cost much less than finding and verifying the addresses. That makes the order useful: screen the candidates first, then spend on the people who fit. The full sample cost $0.07980 in AR data fees. Your agent’s model or subscription fees are separate. These are settled charges from this sample, not a standing quote for future lookups.
| Step | Calls | Actual charge |
|---|---|---|
| Find people at the company | 1 | $0.0005 |
| Check current LinkedIn roles | 2 | $0.0058 |
| Find work emails | 2 | $0.0490 |
| Verify returned emails | 2 | $0.02450 |
| Total | 7 | $0.07980 |
Try it with your agent
Connect AgentRouter once, then give your agent the company domain, the roles you want and a total data budget. Ask for sources and exception flags in the deliverable. When starting with an industry rather than a named company, add a company-discovery step and budget for it separately.
For a useful CSV, keep one row per person: name, company domain, current role, profile source, returned email, finder provenance, verification status and run IDs. The prompt below requests research only. Review the list before using it for outreach.
Using AgentRouter, find up to five marketing or growth contacts at linear.app. Read the current profiles of the two best matches. Find their work emails and verify each returned address with a separate verification call. Keep inferred, catch-all, unknown and invalid results clearly labeled. Return a CSV with role sources, verification flags and run IDs. Keep AR data fees under $0.15 and show the actual total. Do not send any messages.