Four jobs a recruiter does by hand, on every role, every week. None of them wins a placement on its own; each of them loses one when it slips. Hamrr does all four from data already on the desk, and holds the result for a yes.
The same four run on every desk in the office, the same way, and every result is filed against the person who approved it.
Working out who has gone quiet, and then writing each of them something, is the first hour of most desks. Hamrr measures silence per person against the stage they are actually in — three days at interview is not three days at sourced — reads the existing thread so the message answers what was last said, and stacks the results in one queue across every role.
You read and press send. A sequence fires on a schedule whatever has happened; this decides per person, per day, on what has.
An hour a day, per consultant, that nobody has to be reminded to spend.
Each draft carries the silence clock, the stage, the last contact date and the message itself. Discard is one click; nothing is sent without approval unless that category has been handed over.
Most desks lose placements to neglect rather than to a shortage of candidates. Hamrr measures each open role against the benchmark for its own seniority band — not one number for the whole desk — and where a role is short it matches people already in your pool, scores the fit, and checks whether they have since taken a job before offering them at all.
Sourcing tools find you more people. This works the people you have.
Return on a database licence you are already paying for.
The role is measured, the shortfall named, and the matches listed with a fit score and the date each was last contacted. Nobody is emailed until a row is ticked.
Notice period, current salary, right to work, a phone number. Nobody fills them in at the time and everybody needs them at submission, so the gap is found on the day it costs something. Hamrr reads what the ATS already holds — including the per-account custom fields most integrations skip — and lists exactly who is short of what, per role.
It writes only into a field that is still empty, so it cannot overwrite a colleague's entry. Every write is logged against the person who approved it.
Data quality that holds without a manager auditing records by hand.
A completeness score tells you there is a problem. A list tells you whose. Each row states how many fields are missing and whether the gap is blocking a submission yet.
Name, email addresses, phone numbers, home address, referees and profile links come off before a CV reaches a client. The text is removed rather than covered over — a white box still hands the address to anyone who copies the text out — and the space it leaves is closed up, so the document does not read as though something has been torn from it.
It runs by itself when the submission panel opens and the send waits for it, so the safe outcome is not the one you have to remember. Each row states what came off.
The same standard on a new starter's first submission as on your best consultant's.
Three CVs marked client-safe, each naming what was removed, above a covering email addressed to the manager by name. A candidate with no CV on file blocks the send rather than quietly dropping out of it.
A candidate's silence gets chased. A hiring manager's rarely does — there is no field for it, so nothing measures it. Hamrr watches the manager side of every open role the same way: how long since you last chased them, and what they are actually sitting on.
What they owe is derived from where the people blocked behind them are stuck, not guessed at — and it decides the clock, because an offer sign-off and a set of interview notes are not the same wait.
Booking follows the same rule in the other direction: slots stay locked until the client has asked to see someone.
Each wait is read from where the people blocked behind it are stuck, so the manager is chased on the thing they actually owe.
| An offer sign-off | 1 day |
| Interview feedback | 3 days |
| A shortlist to review | 3 days |
| Nothing readable | Standard clock |
Every workflow leaves the same record: what was raised, what was sent, and who approved it. A desk can be read back over any period without anyone stopping to write it up.
Roles running short, people stuck in a stage and clients left unchased are the same three lists for whoever reads them.
A review starts from the record rather than from recollection.
What was sent, to whom, from which workflow, and whether you approved it or Hamrr did it on your say-so. Every send is filed with its full content.
Every workflow above prepares the work and stops. What Hamrr may finish on its own is set in four separate categories, each independently on ask me or automatic, and all four can stay on ask me indefinitely.
Hand over the mildest first — filing notes to your own ATS, where no candidate or client sees the result. Anything done unattended is logged with the full content of what was sent.
Deletions and bulk changes are not delegable at all, in any setting. Each category is set per account, and switching one to automatic takes a confirmation naming what will start happening without you. The four categories, and what each one covers →
Connect a read-only ATS credential and review the first drafts in the same session.
hamrr