Guide · The future of HR
Aligning HR and Operations
for AI transformation
AI is not primarily a technology rollout. It is a redistribution of work, and work is the one thing HR and Operations have never fully agreed on who owns. This guide lays out where the two functions actually collide during AI adoption, the conversations most teams schedule around, and a four-step framework for running the change together.
Why the future of HR runs through Operations
For most of the last two decades, HR and Ops could work in parallel. Ops optimised process; HR managed the people inside it. AI collapses that separation, because the unit of change is no longer a department or a system, and it is a task. And tasks sit inside jobs.
The moment an organisation automates a task, it has quietly redesigned a role. If HR finds out at launch, it inherits a change it had no hand in shaping: unclear job descriptions, managers who cannot answer their team's questions, and a skills plan written after the fact. If Ops runs it alone, adoption stalls for reasons that never appear in the process map.
That is the real future-of-HR question: not which tools HR buys, but whether HR is in the room when operational decisions reshape the workforce.
Where the ownership line actually sits
Deciding what gets automated
- Operations owns
- Owns the process map, cycle times, and the business case for change.
- HR owns
- Owns the role impact, skills gap, and the redeployment plan.
Friction point: Ops picks the pilot before HR knows which jobs it touches.
Measuring whether it worked
- Operations owns
- Throughput, error rate, cost per transaction.
- HR owns
- Retention, engagement, internal mobility, manager load.
Friction point: Two scorecards, one project. Nobody agrees what success looked like.
Owning workforce data
- Operations owns
- Wants operational telemetry on how work actually flows.
- HR owns
- Owns the consent, privacy, and fairness obligations on that same data.
Friction point: The data gets collected first and governed later.
Communicating change
- Operations owns
- Talks in rollout milestones and system cutovers.
- HR owns
- Talks in job security, expectations, and manager readiness.
Friction point: Employees hear the milestone before they hear what it means for them.
The four uncomfortable conversations
These are the questions that get deferred until they become incidents. Put them on the agenda of the first joint planning session.
1. Which roles does this pilot change, and who tells those people?
The single most avoided question in AI adoption. Answer it before the pilot, not after.
2. If this frees up 20% of a team's time, what is that time for?
Unnamed capacity gets reabsorbed into busywork, or read as a headcount threat.
3. What are we explicitly not automating, and why?
A stated boundary does more for trust than any reassurance memo.
4. Who is accountable when the model is wrong?
If the answer is 'the tool', the answer is wrong. Name a human owner.
A four-step alignment framework
01
Map the work, not the org chart
Run one joint session where Ops walks the process end to end and HR annotates every step with the role that performs it. The overlap is your scope; the gaps are your risk register.
02
Agree on one shared scorecard
Pick three operational metrics and three people metrics before kickoff. A pilot that improves throughput while burning out the team is a failed pilot. The scorecard should be able to say so.
03
Name the redeployment path first
Before automating a task, write down where the freed capacity goes: higher-judgment work, backlog, or a named new role. Skipping this step is what turns AI adoption into a trust problem.
04
Run a joint retro in the open
Thirty days in, HR and Ops present one set of findings together, including what went badly. Shared authorship is what stops the two functions from narrating different versions of the same project.
Where to start this week
- Pick the AI pilot furthest along and list every role it touches.
- Book one 60-minute joint session with an Ops lead and an HR business partner.
- Write the shared scorecard: three operational metrics, three people metrics.
- Answer the four questions above in writing, and circulate the answers.
Hear this debated on the show
Episode 001 takes on exactly this question with two guests who have run it in the wild.