Skip to content
Team Nexa
All posts

Success storyIllustrative

How a 400-person health-tech company saw attrition risk building in one engineering team

Exit interviews kept saying the same thing, after people had already left. The People agent saw the pattern at team level while there was still time to act, without looking at anyone's personal data.

3 min read

  • People
  • Product & Tech
  • Strategy

Illustrative story: a composite scenario built on sample data to show how the agents work. It is not a customer testimonial.

The challenge

Engineering attrition across the company sat close to the industry average, so it rarely came up in leadership meetings. The average hid one team. The platform group, which kept the clinical integrations running, lost four engineers it wanted to keep in six months.

Each exit interview mentioned workload and pay. By the time anyone read them side by side, the people were gone, and the engineers who stayed were covering their on-call shifts.

What they set up

  • The People agent read the HRIS and compensation bands read-only, and looked at tenure, pay against band and voluntary exits by team. It never produced a list of names or a score for any individual.
  • The Product & Tech agent read on-call schedules and incident history, and reported on-call hours and after-hours pages as team averages.
  • Before any source was connected, the CHRO and CTO agreed the rules: team-level figures only, small teams reported together with their department so no one could be singled out, and no access to messages, calendars or anyone's activity.
  • The Strategy agent added a line to the Monday brief whenever a team crossed a risk threshold the CHRO had set.

What changed

regretted exits from the platform team, two quarters before and after
4 → 1regretted exits from the platform team, two quarters before and after
on-call hours per engineer once the rotation was widened
−35%on-call hours per engineer once the rotation was widened
team-level attrition review, instead of reading exit interviews after the fact
Monthlyteam-level attrition review, instead of reading exit interviews after the fact

Leadership made two changes. Pay for roles below band was raised to it in an off-cycle adjustment, a quarter earlier than the normal review. The CTO widened the on-call rotation to two adjacent teams and backfilled the open seats first. The agents showed where the pressure was and what each change would cost. People made every decision, and the conversations with engineers stayed with their managers.

The lesson

Attrition signals only help if employees trust how they are made. Keeping the work at team level, with rules agreed before the first connection, meant the CHRO could share the findings openly with engineering managers instead of sitting on a quiet report nobody wanted to discuss.

Keep reading

All posts
  • Use case

    From brief to action: how the agents handle follow-through

    A good brief tells you what changed. Then someone still has to send the email, update the plan and chase the owner. Here is how the agents take that part off your plate, with your approval at every step.

    • Strategy
    • Finance
    • Revenue
    • Operations

    2 min read

  • Use case

    Linking the product roadmap to revenue and cost

    Roadmap reviews talk about features. Board meetings talk about revenue. The Product & Tech agent connects the two, so you can see what each big bet is worth and what it costs.

    • Product & Tech
    • Revenue
    • Finance

    2 min read