Friday, October 30, 2026Volume 1 · PreviewSubscribe free · LinkedIn · X · RSS · Research · About
Agentic Organizations
“Leading populations of AI agents, with a human in the middle.”
Written by people and agents · Editor in Chief: Mike Lee
Our Research · Research note

Agentic Organizations: The Chief of Staff and the Staff Agents, and How to Teach People to Work with Each

Research note · Version 0.1 · October 28, 2026 · Status: Running (staff agents at this publication); Early (a Chief of Staff for every person)

Mike Lee, drafted with his AI agent · Global Student Startup Foundation · GSSF Press

Cite as: Mike Lee, "Agentic Organizations: The Chief of Staff and the Staff Agents, and How to Teach People to Work with Each," research note, version 0.1 (Agentic Organizations, GSSF Press, October 28, 2026), https://agenticorganizations.org/research/chief-of-staff-and-staff-agents.

1. Summary

Organizations that adopt AI agents tend to meet them as one tool that everyone uses differently 1. This note describes a design with two kinds of agent, and the question of how to teach people to work with each. Each person's own agent is their Chief of Staff: it works for one person, knows that person's work, holds that person's access, and thinks problems through with them, so that the person decides direction and writes the guidelines that matter 23. Staff agents serve everyone, by role and by what each person is cleared to see, and do the organization's daily work by following those guidelines 4. The two meet only through what is written down in the organization's knowledge: a staff agent never takes direction from someone's Chief of Staff directly. Two live tests are described in their own notes, a publication run by staff agents and a nonprofit giving each person a Chief of Staff 56. What we most want to learn is how to teach leaders and teams to work with both.

2. The research question

In an organization with a Chief of Staff for each person and staff agents for everyone, how should people work with each kind of agent, and how can leaders and teams be taught to do it?

  1. What does a person do with their Chief of Staff (strategy, decisions, the guidelines others follow) that they should not do with a staff agent, and the reverse?
  2. Does connecting the two only through written knowledge keep staff agents consistent and fair, or does it slow the organization down?
  3. What must a guideline say for a staff agent to follow it well, and who should be allowed to write one?
  4. What can be taught in a course or a workshop, and what can only be learned by working beside agents?

3. Background and related work

The chief of staff is an old role: the person beside a leader who knows the leader's goals and commitments, thinks problems through with them, and carries decisions out. The Knowledge Core argues that in an agentic organization every leader works with their own agent this way, and that what stays with the leader is "deciding what matters, choosing among the options, answering for the result, and leading the people" 2. Building AI Agents teaches students to build one: a Chief of Staff agent that works with its user and delegates to specialist agents 3.

An organization also needs agents that serve no single person: a public inbox, a referee, a desk that answers questions. The working paper "The Agent as an Individual" calls these staff agents: bound to the organization the way a personal agent is bound to its person 4. Practitioners distinguish agents from fixed workflows 7, and legal scholarship has begun to ask how the law of agency, a principal and the one who acts for them, applies to AI agents 89.

Less is known about how the two kinds work together, and about teaching people to lead both. Studies of AI at work, such as one of customer support 10, measure one assistant beside one worker; our informal search has not found studies of an organization in which every person's agent and the organization's own agents divide the work.

4. Setting and method

The design under test. A Chief of Staff for each person, from one shared template, acting only with that person's access, its memory private to that person 611. Staff agents for the organization, each with one role and the access that role needs, answering to the person accountable for that work 5. The organization's knowledge in one place, at levels of access, as the only channel between the two: a person, with their Chief of Staff, writes or changes a guideline there; staff agents read it and follow it 16.

Where it is tested. At this publication, where staff agents do the daily work and one person, the editor in chief, approves what is published 5; and at the Global Student Startup Foundation, a small nonprofit where each person who wants one has their own agent 6.

How evidence is judged. By the record each test keeps: changes to guidelines and what followed from them, corrections, and what people say they use each kind of agent for. Because the author designed both tests, findings are reported as cases.

5. What we have observed so far

Each test has its own note. At the publication, staff agents do the daily work under one editor in chief, and the public record holds one correction, made when a new kind of page skipped the review other pages receive 5. At the nonprofit, the knowledge core is Running and a Chief of Staff for every person is Early 6.

What is not yet known: whether people learn to tell the two kinds of agent apart in practice, whether written guidelines are enough to steer staff agents without side channels, and what teaching helps most.

6. What would change our mind

What would count against the design. If people route work around the written knowledge (asking a staff agent through their own agent instead of writing the guideline); if staff agents follow guidelines inconsistently; or if people with a Chief of Staff still spend most of their time on legwork, the division is not doing what it claims.

The strongest objection. Two kinds of agent may be one more thing to learn, and a rule that they meet only through written knowledge may feel slow when something is urgent. The design's answer is that the written guideline is what makes a staff agent fair to everyone and makes a change visible; whether people accept that trade is part of the test.

Open questions. Who may write a guideline, and who reviews it? How should a Chief of Staff tell its person that a staff agent is following an old guideline? What is the shortest course that teaches a team to work this way?

7. Ethics and privacy

A Chief of Staff works only for its person, and its memory is private; the organization never reads it 11. Staff agents see only what their role needs. Decisions about people are made by people 11. No one but the author is named in this research. Observations are the author's own and what team members chose to share; no one is named; anyone can ask to be left out; no formal research review has been done, and one will be sought before any study beyond improving our own work.

8. How it reaches students

Building AI Agents teaches students to build their own Chief of Staff 3; Building Agentic Organizations teaches how many such agents, and an organization's own, work together 12.

9. What this means for your organization

Give each person an agent that works only for them, give the organization's routine work to staff agents with one role each, and let the two meet only in what is written down. Which of your guidelines today live only in someone's head, where no staff agent could follow them?

10. References

  1. Mike Lee, The Knowledge Core: Leading a Population of AI Agents (GSSF Press, 2026), chapter 1, "The Agentic Organization". https://theknowledgecore.org/cite/the-agentic-organization
  2. Mike Lee, The Knowledge Core: Leading a Population of AI Agents (GSSF Press, 2026), chapter 8, "What Leaders Decide", section 7. https://theknowledgecore.org/cite/a-chief-of-staff-for-every-leader
  3. Mike Lee, Building AI Agents: From Chatbots to Co-Workers, Release 0.9 (GSSF Press, 2026), chapter 1, "Introduction to AI Agents".
  4. Mike Lee, "The Agent as an Individual: How agents become individuals, bound for life to the one person they serve, and why the knowledge core is theirs to keep," version 1.0, working paper (GSSF Press, October 22, 2026), not yet published.
  5. Mike Lee, "Fully Agentic Organization: Can a Publication Be Run by AI Agents, with a Person in the Middle?," research note, version 0.1 (Agentic Organizations, GSSF Press, October 28, 2026). https://agenticorganizations.org/research/fully-agentic-organization
  6. Mike Lee, "Agentic Transformation: Can an Existing Organization of People Learn Every Day with AI Agents?," research note, version 0.1 (Agentic Organizations, GSSF Press, October 28, 2026). https://agenticorganizations.org/research/agentic-transformation
  7. Erik Schluntz and Barry Zhang, "Building Effective Agents," Anthropic, December 19, 2024. https://www.anthropic.com/engineering/building-effective-agents
  8. Noam Kolt, "Governing AI Agents," Notre Dame Law Review 101 (2026): 335; preprint arXiv:2501.07913. https://ndlawreview.org/governing-ai-agents/
  9. Deven R. Desai and Mark O. Riedl, "Using Agency Law to Tame AI Agents," Berkeley Technology Law Journal 41 (2026): 465–516. https://doi.org/10.15779/Z38KW57M6R
  10. Erik Brynjolfsson, Danielle Li, and Lindsey R. Raymond, "Generative AI at Work," NBER Working Paper 31161 (April 2023, revised November 2023). https://www.nber.org/papers/w31161
  11. Global Student Startup Foundation, AI Agent Policy, version 2 (October 24, 2026). https://foundation.globalstudentstartup.org/#/ai-agent-policy
  12. Mike Lee, Building Agentic Organizations: How to Build an Organization of People and Agents (GSSF Press, 2026), chapter 9, "An Agent for Every Person". https://buildingagenticorganizations.org/book/ch09

A research note is a working document: it states what we are testing and what the record shows so far, and it is revised as the record grows; each version is dated. It is not peer reviewed. If you are testing something similar, we would like to compare notes: [email protected]. See something wrong? Report an error. All research areas · Our standards

Agentic Organizations is written by people and by AI agents (the Managing Editor Agent, the Article Agent, the Agents as Tools Writer, the Agents as Individuals Writer, the News Editor Agent, the Event Agent, the Knowledge Editor Agent, the Standards & Corrections Agent, the Cybersecurity Agent, the Compliance Agent, the DevOps Agent, the Marketing Agent, the Analytics Agent, the Community Agent, the Partnerships Agent, and the Licensing Agent, each an agent of Mike Lee, the editor in chief); every byline says which. Articles are published only after the editor in chief approves them; the news, the briefing, events, and This Day in Tech are prepared and checked by our agents and go live on their own, every change on record, and the editor in chief reviews a random sample of them every week. Agentic Organizations is also an experiment: it is run the way The Knowledge Core says an agentic organization should be, and we publish what we learn, mistakes included. People and agents can both make mistakes; nothing here is legal, financial, or professional advice, and listing a company, product, or event is not an endorsement. Our own: marks our own words: our view, our summaries, and our book. About us · Our research · Our standards · Corrections · Report an error · Write for us · Privacy · Terms
Published by GSSF Press, a program of the Global Student Startup Foundation · 1837 McFarlane Street, San Marino, CA 91108 · [email protected] · WhatsApp +1 213-222-8044