Kevin Welter.
ENDE

AI employee in testing

Kai makes your AI team manageable.

Kai is still being tested in my own business. He is not currently available as a finished AI employee or package. I expect to make him available once I have tested him thoroughly and can hand him over properly.

I give Kai a clear leadership assignment. He then reads the results from my AI employees, checks them against my assignment and the actual state of the work, gives targeted feedback and turns everything into one document containing the decisions that remain with me.

Join the communityRead first: Learning Claude
View community

Overview

cases instead of message chaos

Context

risks and decisions stay visible

Kevin

keeps every approval

Manager

How Kai takes on leadership work

Kai does not act on his own judgment alone. My assignment is always the starting point: it defines what Kai should watch, which result he should review and the boundaries within which he can give employees feedback.

Several

employee results brought together

1 assignment

sets the goal and review standard

1 document

condenses decisions for me

Kai brings together reports, protocols, commits and questions from the team. He compares those claims with the actual state of the work and my assignment. If evidence is missing or a result misses the target, he gives the responsible employee feedback or asks a question. Only the points that need my decision go into the final decision document.

One example from the tests so far: Kai spotted an incomplete correction pass, an open permission question and an unverified change across the work of several employees. At the same time, he confirmed that another rollout was fully live. That is the difference from a dashboard: Kai does not merely count activity, he evaluates it within his assignment.

Kai for your business

You define the leadership assignment. Kai keeps the team aligned.

Kai gets no general authority over other employees. You define the goal, review standard, boundaries and the decisions that must end up with you.

How I use him

Bring together the work of several AI employees

Kai checks results against my assignment, gives feedback to the responsible employees and condenses open points into a decision document for me.

Adapted to you

Your roles, your assignment, your escalations

You define which employees Kai coordinates, what counts as evidence, when he may give feedback and which questions must always be escalated to you.

What you change

  • Team and responsibilities
  • Goal and review standard
  • Feedback and escalation rules
  • Decision-document format

Other possible uses

  • Content production
  • Campaign coordination
  • Website rollout
  • Recurring quality control

The process stays: Start from the human assignment, verify results against actual state, give traceable feedback and leave decisions with the person in charge.

What Kai uses to lead the team

Kai does not lead by instinct. His assignment and the documented work results form the basis together:

01

Kevin's assignment

The assignment defines the goal, priorities, review criteria and boundaries. Kai measures results and his feedback against it.

02

The team's evidence of work

Reports, protocols, git history and the actual files show what was really completed and what has only been claimed.

03

Feedback loop

Kai sends questions and specific guidance to the responsible employee when evidence is missing or the result deviates from the assignment.

04

Decision document

At the end, Kai condenses status, contradictions, risks and open decisions so I do not have to search every folder myself.

What he can't do, where I approve

He decides nothing on substance, approves nothing, and commits nothing.

He doesn't change other employees' work, only his own review protocols and mailbox notes.

He doesn't write other employees' personnel files, that isn't his job.

He only hands out substantive tasks to other employees on my explicit instruction. His feedback and questions must stay within the leadership assignment I gave him.

If an approval line or a named file is missing, he writes "not verified" instead of quietly smoothing it over. I stay the only approver, for Kai just like for everyone else.

The current status: Kai is still being tested

Standard

Today: build and test it yourself

The entry point is Learning Claude, plus Hiring an AI Employee for the process and Creating an AI Employee for the build. What's special about Kai: he watches git, reports and sessions from several other employees and checks claims against evidence, instead of just reporting them. This role assumes you already have several employees running, Kai isn't a beginner role.

  • Understand the build step by step
  • Adapt rules and examples to your own business

Premium

Expected to become available later

There is no finished Kai package today. I am still trying the role in my own business and testing whether it works reliably in day-to-day operations. Once Kai has been tested thoroughly and I can document and hand him over properly, I expect to make him available to others as well.

  • Use the documented build as your starting point
  • Define scope, access and approvals for your setup
Join the community

Frequently asked questions about Kai

How do I keep track when several AI agents are working at the same time?

I give Kai a leadership assignment with a goal and review criteria. He reads the employees' results, compares reports and protocols with the actual state of the work, gives feedback when something differs, and writes one document containing the decisions that genuinely remain with me.

Do you need a leadership role for several AI employees, or is a dashboard enough?

A dashboard shows numbers, but it doesn't check whether a claim is true. Kai compares what a report says with what's actually on the file system, and flags gaps instead of smoothing them over. Past a certain number of running employees, just watching isn't enough anymore, and a dedicated review role starts to pay off.

How do you know if an AI agent really did what its report says?

That's exactly what Kai checks for me: he compares an employee's commits, reports and protocols against the real state of the file system and the live website. If something doesn't match, say a missing approval line or an unverified change, he explicitly writes "not verified" instead of quietly closing the gap.

At how many AI employees does a dedicated coordination role pay off?

I can't back a fixed number, I simply have too few data points of my own. In my case, the workforce had already been running for a while before Kai joined. My own rule of thumb: once you stop reading every report yourself, you need someone who does that for you.

What can a leadership role for AI employees decide, and what can't it?

With Kai, the line is clear: he decides nothing on substance, approves nothing, and commits nothing. He may send feedback and questions to other employees within my leadership assignment. He does not set new substantive goals or assignments on his own. I remain the approver.

Several AI employees, one overview, no black box

Kai is not yet available as a finished AI employee. The learning path for building your own AI workforce is already available in the community.

Join the community

Read on

Also:What Can an AI Lead Role Decide?·Learn Claude: The Path in Five Stages·Hiring an AI Employee: The Process·Creating an AI Employee: How to Start·Claude Subagents, Skills, Agents·My AI Workforce: Fourteen Employees, One Boss·Hire Your First AI Employee in One Hour·Back to the org chart