A People Ops system that keeps learning

My view of what People Ops should look like, and the systems I want to build.

people opsAI agentslearning design

People Ops should learn from the work people actually do, connect capability gaps to the company roadmap, and turn that understanding into better support. This deck sketches how I would build that system, with AI helping run the routines and people making the decisions.

Download the PowerPoint

A People Ops system that keeps learning. Proposed architecture. An encounter. A blocked task or a conversation reveals what someone needs.. A capability gap. Compare work evidence with the company roadmap.. A running system. Workers update records and propose support. People decide.
1 / 7 · A People Ops system that keeps learning
Capability gaps and the company plan. Roadmap need. Capability gap. Company decision. People Ops response. Reliable AI launch. Prototype only.. Live run untested.. Pilot before launch. Arrange coached practice. Customer self-service. Guide exists.. Users still need help.. Test a small cohort. Generate learning for real tasks. Expand the team. Missing expertise. Train, redeploy, or hire. Define profiles and our own evals. Market trends refresh the capability backlog. Leaders agree on roadmap changes.
2 / 7 · Capability gaps and the company plan
Learning content made for the work. Task + personal path. Sources become new content. Human review + practice. A personal path. Choose the next real task and the support this person needs.. Generate and review. Create a short guide, video, or practice task from trusted sources.. Learn in the work. Use it on the task. Review the outcome and update the path.
3 / 7 · Learning content made for the work
A bot team connected to people. Chief of staff. Human People Ops. Review. Learning. Recruiting. Check-in. Routines. Employee check-in. Employee explains a blocker. Helpers report to the check-in worker and chief.. The chief brings requests to People Ops. People respond and follow through.
4 / 7 · A bot team connected to people
Turning People Ops routines into automation. Define the routine. Build the shared engine. Run the workflow with approvals. Payroll. Cutoff triggers input checks. Finance reviews and approves.. Immigration + policy. Due dates trigger document prep. Specialists review exceptions.. Team events. Gather preferences, draft a plan, then let people choose together.
5 / 7 · Turning People Ops routines into automation
Recruitment built around our own evals. Gap + profile. Task + rubric. Find relevant work. Review evidence. Company gap. Roadmap gap defines the role. and profiles to reach.. Company evals. Our own task and rubric.. Quality and ownership.. Collaboration in the work.. Sourcing channels. GitHub, X, portfolios.. Referrals + communities.. Relevant contributions.. Evidence review. Evidence against rubric.. Candidate adds context.. People decide.
6 / 7 · Recruitment built around our own evals
Two proofs of concept, then a company pilot. Build A and B. A: early checks + backup owner.. B: spare-device pool.. Prototype in separate branches.. Compare and choose. Compare A and B with people.. Choose a small internal pilot.. One loop to reuse. Pilot inside the company.. Measure readiness and delays.. Improve the chosen workflow.
7 / 7 · Two proofs of concept, then a company pilot