Hawkeye AI

The AI that lives inside your system of record.

Hawkeye reads across the candidates, clients, jobs, notes, calls, emails and pipeline history your team already creates — then turns the boring half of the desk into something you approve instead of something you type.

Institutional memory

Good recruiting context should compound.

Years of notes, calls, messages and search history are the most valuable thing an agency owns and the least usable. Hawkeye makes that history answerable — for the recruiter who built it and the one who joined last month.

  • 01 Candidate matching grounded in your own database
  • 02 Instant summaries across a relationship, a company or a req
  • 03 Follow-ups and stalled work surfaced before momentum dies
  • 04 Human control over every consequential action
Who should I call today?Which send-outs have gone quiet?Summarize our history with this clientDraft a reconnect note to this candidateFormat this resume for the submissionWhat did we agree on the last call?

Who do we already know for this Operations Manager req?

6 people from your databaseGrounded
1
Avery ReidSpoke 18 days ago · placed nearby in 2024
Why
2
Noa CastilloWarm — referred by an active client
Why
3
Riley OkaforApplied last year · never contacted
Why
From question to action

Designed for the desk, not for a demo.

The impressive part of AI in recruiting is not the sentence it writes. It is whether the thing it suggests survives contact with a real client relationship.

01

Ask in plain language

Search the recruiting graph the way you would ask a colleague, instead of remembering which filter on which screen you used last quarter.

02

See where it came from

Every answer points back at the record behind it — the note, the call, the thread, the stage change. If the evidence is thin, Hawkeye says so rather than padding it.

03

Approve before anything moves

Drafts, summaries and suggestions are free. Sending a message, changing a stage or editing a record is a proposal that waits for a recruiter.

Hawkeye accessMCPSettings → API keys
In the appLauncher on every screen, plus the full Hawkeye page
From ClaudeConnect Poach as an MCP server and work your desk from chat
Scoped keysIssued per user, shown once, revocable from Settings
Same guardrailsDraft-first sequences and do-not-contact apply everywhere
Wherever you work

Your ATS, available to the assistant you already use.

Poach ships an MCP server, so Hawkeye is not trapped behind a login. Point Claude — or any MCP client — at your workspace with a scoped key and ask it to pull a pipeline report, draft outreach or format a resume, with the same permissions and the same guardrails as the app.

See what connects
Honest limits

What Hawkeye will not do.

Every vendor will tell you what their AI can do. This is the other list — and it is the reason recruiters end up trusting the first one.

Limit

It does not guess about people

If the record does not say where someone works, what they earn or whether they are open, Hawkeye omits it. No invented facts, no confidence scores with nothing behind them.

Limit

It does not send on its own

Outreach is drafted and queued. A human starts it — which is also why sequences are created as drafts in the first place.

Limit

It respects the same rules you do

Do-not-contact, stop-on-reply, workspace permissions and communication preferences apply to Hawkeye exactly as they apply to a recruiter.

Limit

It works on your data, for you

Hawkeye reasons over your agency’s workspace. It is an operating layer on your database, not a shared pool of everyone’s candidates.

Hawkeye AI

Give your team a better view of the desk.

See the platform