An AI chief of staff for the CEO's office.
I set up Claude or ChatGPT, whichever your company already uses, as a chief of staff for the CEO: connected to your email, calendar, meeting recorder, CRM and documents, so it knows your business and has a brief ready before every meeting. For CEOs of 50 to 500 person companies in North Carolina: the CEO first, then the leadership team.
What changes in your week
Four rituals, each with the same shape every time, so you can read them in seconds.
-
7am on weekdays
Morning brief
Today's meetings with a one-line brief each, at most five things that need you with why and by when, and anything that has waited on you for more than three days. Ranked by your priorities, not by what arrived last.
-
Every morning, one for each meeting on your calendar
Pre-meeting brief
One page: who you are seeing, what you last agreed, what is open on either side, and what to decide in the room. Built only from your email, meeting transcripts, CRM and the Company Brain.
-
Friday afternoon
Friday review
The week scored against your priorities: what moved and the evidence, what did not and why, what you promised and to whom, and next week's one thing.
-
Whenever you ask
Decision memo
It asks you three questions, one at a time, before it writes anything. Then: your assumptions ranked by how much the decision rests on them, a pre-mortem, the question you have not asked, and a recommendation with its confidence.
It reads; it does not send. Every brief, review and reply lands as a draft or a document until you act on it.
What the CEO Chief of Staff Install includes
- The Company Brain: one document with your priorities, the people who matter, deals in flight, how you decide and your standing rules, built from an interview with you in the first session.
- Read-only connections to Microsoft 365 or Google Workspace, your meeting recorder, your CRM and your documents.
- A source line at the end of every brief that says what it checked and what it could not reach, so it never quietly fills a gap.
- A one-page data-handling answer your IT person can read.
- A handoff to your executive assistant, who owns the Company Brain once it is working.
The method inside
The method inside is Geoff Woods' CRIT (Context, Role, Interview, Task), from The AI-Driven Leader. Most leaders do CRIT by hand in every chat. The install makes each step permanent: Context lives in the Company Brain and the connections, Role is a set of named skills, the Interview is a step the decision memo cannot skip, and the Task runs on a schedule.
The guarantee
I get it working. I keep going until all four rituals run on your own email, calendar and meeting recorder, however long that takes.
Once it is working, it is yours, and your executive assistant runs it from a one-page runbook. If you want me to keep it sharp, the Leadership Operating Rhythm adds your leadership team one executive at a time, with a monthly leadership session, a quarterly reset and a monthly refresh of sources and context, on a 90-day initial term.
I take four installs a month.
Who sets it up
Stephen Thorn, AI Consultant for small businesses in Raleigh-Durham. I currently advise a cohort of CEOs on AI. Former Marine officer, OpenAI Select Partner, 200+ people trained to use AI at work. More about me.
Who is this for?
For CEOs of 50 to 500 person companies in North Carolina: the CEO first, then the leadership team.
Does it work with ChatGPT or only Claude?
Either. It runs on whichever your company already uses: the briefs, the review and the decision memo work the same way in both.
Will it send email or change anything on its own?
It reads; it does not send. Every brief, review and reply lands as a draft or a document until you act on it.
Do Anthropic or OpenAI train on our data?
On Claude for Work, Anthropic says it does not use your inputs or outputs to train its models by default. The exception is a chat someone sends feedback on with the thumbs up or down button (Anthropic Privacy Center, updated August 18, 2026). On ChatGPT Business and Enterprise, OpenAI says it does not train on your business data by default, and you own your inputs and outputs (OpenAI enterprise privacy page).
What happens to conversations we delete in Claude?
On Claude for Work, a conversation you delete leaves your chat history immediately and is deleted from Anthropic's back-end systems within 30 days (Anthropic Privacy Center, updated July 1, 2026).
What does it connect to?
Read-only connections to Microsoft 365 or Google Workspace, your meeting recorder, your CRM and your documents.
What does it cost?
It depends on scope and gets set on a free 20-minute call.
What if it does not work?
I get it working. I keep going until all four rituals run on your own email, calendar and meeting recorder, however long that takes.
What happens once it is working?
Once it is working, it is yours, and your executive assistant runs it from a one-page runbook. If you want me to keep it sharp, the Leadership Operating Rhythm adds your leadership team one executive at a time, with a monthly leadership session, a quarterly reset and a monthly refresh of sources and context, on a 90-day initial term.
Is this made by Anthropic or OpenAI?
It is built by Stephen Thorn and is not made by or affiliated with Anthropic, OpenAI or Geoff Woods.
Want to see what yours would look like?
Book a free 20-minute call. We will look at how your week runs today and whether this fits.