Hi there!

Welcome to the 31st edition of Work in Beta.

In this edition: a Chief of Staff used to be something only CEOs and founders got. We break the role into five jobs and show how to build each one with AI.

So, let’s dive in!

You know the role. A Chief of Staff is the right-hand person for someone at the top, and the whole job is keeping that person ready for whatever the week throws at them. They walk into meetings already briefed. Nothing gets dropped.

And you only got one if you sat near the very top, running a company. Everyone else handled their own mornings, prepped their own meetings, and chased their own follow-ups, week after week.

That's the part that has changed. The jobs a Chief of Staff does are exactly the jobs AI has become good at. You can't hire the person. But you can now build the role.

You don't get a ready-made Chief of Staff. You build yours one job at a time.

A Chief of Staff really does five jobs, and each one is a request you'd hand over: brief me before the day starts, prepare me before meetings, capture what happened, track the to-do list, and help me think through the decisions.

Here's each job, and what stays yours.

A Chief of Staff scans what's coming so the day never arrives as a surprise. The AI version is a morning brief you ask for. Hand AI your calendar, your inbox, your notes, your task list and your chats in tools like Slack. Then ask: what matters today, what changed since yesterday, what needs a reply, which meeting needs preparing?

A good brief highlights the important stuff before 10 a.m. A bad one reads your calendar back to you. Either way, it highlights; you choose what the day should be about. Once the brief is hitting the quality you're happy with, automate it to run on a schedule: ChatGPT, Gemini and Claude all run scheduled tasks now, and Copilot does a version inside Microsoft 365.

This job ensures that nobody senior walks into a room cold. At Upflow, a payments company, the Chief of Staff checked the CEO's agenda and made sure everything was documented before the CEO ever joined the conversation.

The AI version of this is one page you can ask for before each call: who's coming, what happened last time, what's still open, and what's worth asking.

OpenAI's own sales team built this. Reps used to spend close to an hour preparing for a 30-minute call, pulling context out of dozens of systems, and now an assistant builds the brief and each rep gets about a day back every week. The rep still runs the meeting.

After every meeting, someone has to write down what was decided and who owes what by when. If nobody does, the meeting slowly un-happens.

Morgan Stanley built this job for their advisors: with the client's consent, it produces the notes, the action items, a follow-up draft and a record in their client system after each meeting, saving one advisor about 30 minutes per meeting. The advisor still edits and sends. Nothing is sent without human approval.

We run our own version. Our meeting notes collect through the day, and every evening an automated run summarizes them and posts the day's update to our Slack. The run takes 15 to 20 minutes, and we get a clean record of what happened where. It reads like a thoughtful colleague reporting back on the day.

Your task list has a weakness: it only captures what you remembered to write on it. The task you picked up in the middle of running Monday's review never reached the list, because you were busy running the meeting. A Chief of Staff owns these loose ends.

The AI version can read your meetings, your mail, your chats and your notes, and write the list itself.

Our evening summary does this second job for us: from the day's notes we extract the task list, the follow-ups and the potential roadblocks. The AI surfaces and drafts. Whether to chase, and how hard, stays with you.

The most senior job of the five. A good Chief of Staff challenges your thinking and says the things nobody else in the room will. The AI version is a note before you decide: the options, the trade-offs, what's missing, and what to ask next. You still make the call; you just stop preparing for it alone.

We use this job for our writing. When a topic idea lands, we do a quick braindump of the raw thought. Then two agents take it up, one for LinkedIn and one for this newsletter, and go researching. They come back with a document holding everything they found: what's already been said, the links, the gaps, and an honest view on whether the idea holds. We read it and keep what makes sense.

That's the full role. But these five jobs aren't things you set up once and forget. Together they make a system, and a system needs managing. Models get upgraded, and a step that worked for months starts behaving differently. Your own work shifts too, and the brief that fit January doesn't fit August.

Treat it like any hire: review the work every few weeks, and when the output drifts, adjust the instructions. When you find something worth improving, refine one thing at a time. The system should grow the way it started. Slowly.

So where do you begin? The same way you'd onboard a human hire: one job at a time. Hand over the reading jobs first, where a mistake won't be costly. Jobs that write things other people will see should come after you're satisfied with the output for a few weeks. And anything that can't be undone should always need your approval. The tools already work this way: Claude's Gmail connector will draft an email for you, but it will not send it. That's not a missing feature. That's the design.

Pick the job where your week leaks the most. Meetings eat your week? Start with capture. Mornings scattered? Start with the brief. Accountability slipping? Start with the tracker. Give AI one day's material and ask for the output once.

A Chief of Staff used to be something you were given, if you were senior enough. Now it's something you can build. The people at the top never got all five jobs from a new hire in week one either, so don't ask that of yours. Build one job. Let it earn the second.

Most professionals know AI can do more for them. The gap isn’t awareness - it’s knowing where to start, what to change, and how to make it stick.

That’s what we work on through Work in Beta.

For individuals, we run working sessions, not teaching sessions. You bring a real problem from your actual work; we build the solution with you, live. You’re borrowing our learning curve instead of grinding through your own. You walk out with something that works and the muscle to keep going. When you get stuck later, we’re a message away.

You probably have a version of at least one of these:

Bring that. We’ll build it with you.

For organizations, AI adoption is a people problem, not a technology problem. Your teams have the tools, what’s missing is the translation layer between AI capability and daily work. Which processes to redesign, which habits to break, how to build genuine fluency, not just awareness. We help close that gap through hands-on training, process redesign, and deep adoption engagements. Not advisory, forward-deployed.

If any of this resonates, email us at [email protected] and we will figure out how to work together.