The attrition signal was there.
Three months ago.
A passive intelligence layer that reads your existing organisational data, surfacing flight risk, commitment drift and pipeline health, without surveys, new tools, or behaviour change.
People signals surface
after it’s too late to act.
Every voluntary resignation follows a pattern, sentiment shifts in messages, participation drops in meetings, work rhythms change. The signal is visible months before the notice. But nobody reads 10,000 messages a week to find it.
Meanwhile HR burns hours on manual pipeline tracking and status chasing, work a configured system handles on its own. The strategic work, the part only humans can do, never gets the time.
How a signal becomes a saved hire.
It reads what already exists, and never acts on a person without a human approving it.
What lands on the screen.
A representative reconstruction, illustrative, built from the canon, not a live screenshot.
Daily HR brief, surfaced privately to the right manager
Drafts proposed, never auto-sent · self-hosted · figures are design targets
See it on real work.
Concrete scenarios, what goes in, what comes back.
A top performer goes quiet, volume down 60%, two 1:1s skipped.
Flight-risk 0.78 surfaced privately to her manager, with talking points drafted.
Three candidates idle 14+ days at the top of the pipeline.
Stale-candidate alert with follow-up emails and interview slots ready to send.
“I’ll share the JDs by Wednesday”, Wednesday passes, silent.
Commitment flagged as overdue with a gentle nudge, nothing falls through.
The systems we build
for HR & People Operations.
Each is a standalone engagement. Start with one, the systems compose over time.
Organisational AI OS
Passive intelligence across Teams, email, calendar and HR data, flight-risk signals, commitment tracking and a daily HR brief, delivered automatically.
View solution →Rapid Automations
Offer tracking, onboarding coordination, recruitment screening, targeted automations for high-volume HR processes.
View solution →Agentic Pipelines
CV-to-structured-assessment screening pipelines, with an eval harness on scoring consistency, delivered to the hiring manager.
View solution →Being built now,
honestly, in the open.
In active development for its first enterprise client, the figures below are design targets, not measured results.
AI CompMate, HR Layer
Passive intelligence for HR: flight-risk, pipeline health, attrition prevention, in active development for its first enterprise client.
The figures below are design targets for the first deployment, not measured results.
- Target: surface voluntary attrition signals before resignation
- Target: HR recovers 3–4 hours/week of manual status chasing
- Every commitment tracked without manual status chasing
- Self-hosted, people data never leaves your infrastructure
HR & People Operations-specific
engineering decisions.
Domain knowledge baked into the architecture, not discovered during delivery.
Not sure which build
comes first?
Start with the service that maps the problem, then we scope the system that fixes it.
Tell us what you
want to solve.
A 30-minute call. We'll tell you which solution fits your situation, what the build involves, and the eval it would have to pass, engineering discussion from the first call, no sales pitch.