What agentic AI in HR can do
Agentic AI can carry several steps without waiting for another prompt after each one. In my HRMC tests, a source check produces a brief for a person to review. The approved follow-up stays attached to the same case.
The person reviewing the work still decides how a policy applies and what action to take.
For a longer explanation, read what agentic AI in HR means.
The tools have different jobs.
Here is what I use each tool for:
- OpenClaw: runs scheduled source checks and keeps the review queue and run history.
- ChatGPT + Codex: planning, building, testing, and updating applications.
- Claude Code: a second place to write and review code.
- Hermes: earlier experiments. I no longer use it for ongoing work.
OpenClaw is where I have done the most hands-on work. I keep its access limited to the tools and sources needed for the task.
Scheduled work needs visible run history. Otherwise you learn it stopped when Monday's brief never arrives.
01 · Input
Approved source
A policy or compliance update
02 · Agent work
Review packet
Source, summary, missing facts, suggested owner
03 · Human gate
HR reviews
Interpret, question, change, approve, or stop
Agent stops here04 · Follow-through
Named owner
Action, due date, status, and proof it was finished
- 64/64
- repeatable checks passed
- 10/10
- known items found in a small source test
- 240
- local requests in five minutes without a failure
What I check after a correction
I now check whether a reviewer’s correction stays attached after the source changes or the workflow runs again. The case study records the failure and dated local checks.
What you can hire me to do
- HR process review: map one process, find where it breaks, and decide whether a tool would help.
- Workflow redesign: redesign the steps, prepare the information each reviewer needs, and agree on what to test.
- AI pilot review: check the results and correction work before deciding whether to expand, change, or stop the pilot.
I do not build agents that decide discipline, termination, accommodations, leave, investigations, pay, promotions, credibility, or other employment outcomes.
A production deployment still needs the client’s engineering, security, privacy, legal, and business owners. See the workflow method and advisory services for deliverables and limits.
Questions I get about the work
When does it make sense to bring you in?
When an HR process keeps getting stuck and you need help finding the cause or testing a change.
Can OpenClaw be used for HR workflows?
Yes. I use it for HR Mission Control’s scheduled source checks. Work involving company data needs approved access and a security review.
What is a good first agentic AI use case in HR?
Try a weekly HRBP brief using public or made-up notes. Check whether it keeps facts and open questions separate and gives the HRBP enough context to choose the next step.
Is HR Mission Control production software?
No. HR Mission Control is a private prototype tested with public and made-up data.