Back to Insights
KieworkHR TechAI ProductsWorkflow

Inside Kiework: How an AI-First HRMS Decides Who Edits the Workflow

Javad PK
Written byJavad PK
14 April 2026
2 min read
Inside Kiework: How an AI-First HRMS Decides Who Edits the Workflow

The gap we kept hitting

We kept seeing the same pattern across our HR clients. Engineers wanted to ship AI features fast. HR leaders wanted full control over the rules. The two groups spoke different languages, and every change cost a week. Kiework is the AI-first HRMS we built to close that gap.

What Kiework is

Kiework is an AI-first conversational HRMS. Employees and managers do not log into a portal full of forms; they talk to Kiework on the channels they already use, and the system handles leave, payroll, attendance, compliance, and reporting in the background. Underneath the chat surface is a workflow engine that lets HR leaders configure approval flows visually while engineers extend the library of available step types.

Two audiences, one tool

The hard problem in HR tech is that the people who need to change the rules are not engineers, and the people who can change the rules safely are. Kiework solves that with a single internal abstraction: every HR flow is a directed graph of steps, where each step is a model call, a tool call, a condition, or a human approval. HR leaders edit graphs visually; engineers add new step types on the platform side without breaking the flows already in production.

Design decisionKiework choice
Who edits flowsNon-engineers via a visual builder
Where flows liveVersioned JSON in the database
How flows failAlways with a human-readable reason
How flows recoverResume from the last successful step
How flows are testedEval set per flow, run on every save

Workflows that ship in production

  • Leave approvals. Read the policy, check balances, draft a response, route to a manager if anything is ambiguous.
  • Onboarding. Generate offer letters, schedule the first week, surface the right documents to sign.
  • Compliance Q and A. Retrieve the relevant policy passages and answer with citations.
  • Payroll runs and exception handling. Group employees by jurisdiction, run the calculation, escalate exceptions to a human reviewer.

Lesson learned: The single biggest unlock was the human approval step. The HR team stopped fearing AI the day they realised they could put a human checkpoint anywhere in a flow with two clicks.

Where to learn more

Try Kiework at kiework.ai. For the broader agent-runtime ecosystem and how the public projects (OpenClaw, Hermes Agent, MoltWorker) compare, see three one-click paths to a self-hosted AI agent.


Javad PK

Co-founder & CEO

Co-founder and CEO of Kiebot. Writes about engineering leadership, product strategy, and how founders should structure their first engineering team.

View full profile →

Want to ship something like this on your product?

Talk to Javad or another engineer from the team that wrote this. No sales deck, just a real conversation.