LIVE · FLIGHT-RISK SIGNALS
Workforce intelligence · for the first 30 days of silence

Your best people are already telling you
they’re leaving.

On average, 71 days before a resignation lands, the calendar goes quiet, the reviews get shorter, and the 1:1s start slipping. Talent& reads it early — across calendar, comms, HRIS, and review data — and tells you who to talk to, this week.

Book a 30-min read of your org → See the demo first
30 min to first insight Read-only connectors SOC 2 architecture · GDPR
SAMPLE ORG · 850 FTE · Q2 LIVE
Stable 779
Watch 42
Flight risk 29
FIG.01 · Flight-risk distribution across a sample 850-person org
The 71-day window before the
goodbye email

In most orgs, the signal precedes the resignation by ten weeks. Calendar density drops. 1:1s slip. Review activity thins. Manager conversations get shorter. By the time HR sees the letter, the conversation you could have had is already two months gone.

TALENT& RESEARCH · COHORT STUDY, N=11,400 DEPARTURES · 2024–2025
02 · DemoFlight-risk simulator

A sample org.
Click anyone.

Six real-ish employees from a fictional 850-person company. Click a person to see what Talent& sees — the signals that moved them onto the flight-risk list, and what it costs the org if they actually leave.

Click a person
to run the simulation.
This demo runs on fake data. We can run the same read on your real org in 30 minutes.
Book a 30-min read →
03 · The three signalsWhat we watch

Three signals.
One instrument.

We measure the work, not the worker. No surveys, no sentiment scoring, no keystroke tracking. Read-only connectors to the systems you already run.

What we read. Calendar density, 1:1 cadence, meeting decline latency, code-review throughput, document-share patterns, internal-mobility clicks, PTO shape, after-hours drift. Nineteen features in total, weighted to your org’s past departures.

What you do with it. A weekly list of names, ranked, with the three specific things that moved them up. Your managers have a grounded reason to reach out. Our customers keep roughly 1 in 3 of the people we flag — the ones they didn’t know were at risk.

EXAMPLE · R-0429
Martín Okafor · Staff Eng, Platform
  • 0.82 risk · up from 0.68 in seven days
  • 3 1:1s cancelled, each within 90 seconds of the invite
  • −41% code-review throughput over four weeks
  • −24% calendar density
Action → Skip-level 1:1 with Martín this week.

What we read. Unsolicited peer mentions in documents and PRs, cross-team collaboration density, scope of work absorbed after departures, review quality, mentor signal from junior’s growth curves, revision counts on strategy docs.

What you do with it. A list of the people your competitors will recruit next if you don’t promote, scope-up, or compensate. We see them roughly 90 days before they show up on an external recruiter’s LinkedIn search.

EXAMPLE · G-0221
Aanya Patel · Sr. PM, Growth
  • Peer mentions +180% in fourteen days
  • Quietly absorbed the scope of two departures
  • Collaboration density in the 94th percentile
  • Has not asked for a title change
Action → Promote, scope-up, or brace for a counter-offer.

What we read. Cross-team collaboration density, manager-skip cadence, DM-to-channel ratio, meeting-network fragmentation, on-call rotation fairness, handoff latency. Measured at team level only — we do not score individuals on culture.

What you do with it. An early warning when a team is quietly unbundling. The pattern that preceded your last attrition cluster is the same pattern we’ll flag on the next one, six weeks early.

EXAMPLE · H-1104
Sales · West · 24 reps, 3 managers
  • Collaboration density −18% over four weeks
  • 3 reps haven’t met their manager this quarter
  • Channel activity redistributed to DMs
  • Matches the Q3 cluster pattern
Action → Managers meet the three reps this week. Reopen the West channel.
Which of these would matter most for your org next quarter? We’ll show you on a call.
Book a 30-min read →
04 · LiveSignals feed

What Talent& surfaced
for one customer, today.

A sample of the actual signal cards a People team at a 900-person company opened this morning. Names changed. Numbers real.

05 · MethodHow we read

Read-only.
Thirty minutes to first insight.

No surveys. No laptop agents. No keystroke tracking. We measure the work, not the worker — at team level for health, at person level only for flight-risk and gem detection, and only on signals already inside the systems you already run.

Inputs
  • HRIS · roles, tenure, comp bands
  • Calendar · cadence, density, decline latency
  • Comms metadata · Slack, M365, Workspace
  • Code review · GitHub throughput & mentions
  • Docs metadata · authorship, share-out, revisions
Metadata only. We never read message contents.
Outputs
  • Weekly flight-risk list, with three specific reasons per name
  • Hidden-gem watchlist, ranked by rise rate
  • Team-health map, flagged six weeks before drift becomes attrition
  • Quarterly boardroom report — exposure modeled in dollars
Delivered to the one person at your company who should have it.
Posture
  • Read-only connectors. We never write to your systems.
  • Team-level only for culture & health signals.
  • Every score is explainable. No black boxes.
  • Employees can see their own file on request.
  • SOC 2 architecture. GDPR. Data residency on request.
Ethics posture: published, versioned, signed by our team.
06 · CallBook a read

Thirty minutes.
Your org. Our read.

Bring a laptop. We’ll stand up read-only connectors during the call and walk your team through the first flight-risk list, the first hidden gems, and the first team-health map — live, on your actual data. You keep the report either way.

I.

You’ll walk away with

  • A ranked list of your 20 most at-risk people, with the three signals driving each name
  • A shortlist of the people in your org most likely to leave for a competitor in 90 days
  • A map of which teams are quietly drifting
  • An exposure model in dollars, signed by our analyst
II.

Who should be in the room

  • Chief People Officer or VP People Analytics
  • Someone from IT who can authorize read-only OAuth
  • Optionally, your CFO — they tend to stay for the exposure slide
III.

Pick a slot

Not ready for a call? Email us your org size — we’ll send a sample report