If your team meets in Google Meet, the useful part is not the transcript itself. It is the list that comes out of it: who is doing what, by when, and what is still undecided. Here is the workflow I teach, from starting the transcript to a follow-up email, using ChatGPT.

One note first. Transcripts in Meet depend on your Google Workspace plan and may need your admin to turn them on. If you do not see the option, that is usually why.

Step 1: Start the transcript in the meeting

In the meeting, open the Activities panel and choose Transcripts, then start the transcript. A badge shows that it is on. Tell people it is being transcribed.

When you stop it or the meeting ends, Meet saves the transcript as a Google Doc in the host's Meet Recordings folder in Google Drive, and typically emails a link to the host.

Step 2: Copy it from the Google Doc

Open the Doc. Click Edit, then Select all, then Edit and Copy. That is Ctrl+A and Ctrl+C if you prefer the keyboard.

Step 3: Tell ChatGPT what you want before you paste

This is where most people lose the value. If you paste a transcript with no request, you get a vague recap. Open ChatGPT, click the message box and send the request first, then paste the transcript below it. Press Shift+Enter for a new line, because plain Enter sends the message.

You're my operations assistant. Below is the transcript of today's team meeting. Give me:
1. A three-sentence summary
2. The decisions we made
3. The action items as a table: task, owner, due date
4. Open questions
Only use what's in the transcript. If an owner or date isn't said, write "Not stated".

The last line is doing the real work. It stops ChatGPT from guessing an owner or a date nobody said out loud.

What came back in the practice meeting

In the free lesson I use a made-up bakery meeting, Harbor Street Bakery, with four people, a lot of back and forth and a few ums. The result was a short summary, three decisions, four action items and two open questions.

Two of the due dates said Not stated. That is correct. Nobody said when the library catering quote or the oven repair call would happen. This is the point where a lot of people go wrong: the summary looks finished, and they send it.

Step 4: Check it against what you know

You were in the meeting, so check the table against your memory. Read the decisions. Look at every owner. Look at every date that says Not stated and fill it in yourself. Names get misspelled in transcripts and a task can land on the wrong person, so this is a read, not a skim.

Then fill the gaps in the same chat, in one message:

The library quote is due Wednesday, and Luis will call about the oven today.

Step 5: Ask for the follow-up email in the same chat

ChatGPT still has the whole meeting in this chat, so you do not paste anything again. Ask for a short, friendly follow-up email to the team with the decisions and the action items. Read it once, change anything that does not sound like you, and paste it into your email.

If the transcript comes as a file instead, say a Word document someone sent you, you can upload it with the plus menu instead of pasting. The lesson covers that route too.

What I leave out

Before I paste, I take out personal side conversations, names where roles will do, contact details and any private client or HR detail. I add the transcript to the current chat as task input, not to memory or custom instructions, and I stay within my company's AI rules.

A table with a confident date in it can still be wrong, so the check in step 4 is not optional.

Where the table goes next

Paste the checked table into a Google Sheet with a Status column and you have a simple tracker. If you want that automated, there is a separate lesson that takes a recap into tracker rows with n8n. For a client call, there is an accounting-firm version that ends in a follow-up email to the client.

Try the whole thing as practice

The free lesson, Summarize a Meeting Transcript with ChatGPT, has 24 steps and takes about nine minutes. You do each step yourself in a practice copy of ChatGPT, it checks your work, and nothing touches your real accounts. It starts at the Google Doc and ends with the follow-up email.

Read the full Google Meet guide: Google Meet Transcript to Action Items with ChatGPT

Someone on your team pastes a client spreadsheet into a free chatbot on a personal login. Someone else pays for an AI note-taker on a company card and records every client call. A draft email with a made-up deadline goes out because nobody read it closely. None of it is malicious. People are trying to get work done faster. The problem is that it happens in private, with tools nobody approved and settings nobody checked.

The answer is not to ban AI. People use it anyway and stop telling you. The answer is a short set of rules that say yes to AI with clear lines. In a 5 to 50 person office that fits on one page and takes an afternoon.

I put the pieces into a free starter kit and a course. Here is the short version. This is practical guidance, not legal advice, so if your work is regulated, have a lawyer or compliance person review your final policy.

The seven pieces

1. Data rules. What must never go into AI?
2. Approved tools. Which tools and accounts may we use?
3. Admin settings. Are the accounts set up safely?
4. Policy. Where are the rules written down?
5. Guardrails and training. Do staff and AI agents know the rules?
6. Human review. Who checks AI work before it counts?
7. Incident plan. What do we do when someone gets it wrong?

1. Start with data rules

Everything else depends on this one. The simplest rule that works is three levels:

• Public: text already on your website, published prices, general questions. Fine in any approved tool.
• Internal: draft memos, meeting notes, process documents. Approved tools on the company account only, with names and account numbers swapped for placeholders like [CLIENT].
• Confidential: passwords, bank and card numbers, Social Security and tax ID numbers, health details, HR records, and anything a client asked you to protect. Never goes in as is.

Put real examples from your own work on the sheet. A quick test: would you be fine if this text showed up in a stranger's chat?

2. Make the approved way the easy way

Shadow AI shrinks when the approved option is simple. Pick one main assistant on a business plan, write it down, and give people an easy way to ask for anything else.

Your approved tools list is just a table: the tool, the plan, who can use it, the highest data level allowed, an owner and a review date. Before anyone signs up for a new tool, ask whether the vendor trains on your data, how long it keeps it, and whether an admin can control sharing.

3. Check the settings on day one

Most leaks come from default settings on accounts nobody configured. Whatever tool you approve, look at these:

• Model training. On personal plans, switch off the setting that lets the vendor use your chats to improve its models. OpenAI says its Business and Enterprise plans do not train on business data by default.
• Shared links and memory. Anyone with a shared link can read that chat, and memory is risky on a login several people share.
• Who is on the account. One company workspace, no client work on personal logins, and remove people the day they leave.

4. Write the one-page policy

Keep it under a page, in plain words, with names. It should cover approved tools only, the never-paste list, what is fine to do (so people are not afraid to use AI at all), the settings the owner keeps in place, and a rule that a person checks every output before it reaches a client, a vendor or the public. Automations and agents that send, pay, delete or post need approval first.

Add two more things: an owner with a review date, and a same-day, no-blame rule for mistakes. That last one is the rule firms forget, and it is what lets you fix a problem while you still can. You make the decisions; AI can do the drafting. I use Claude to turn seven rules into a one-page policy, an announcement email from the boss and a five-question staff quiz.

5 and 6. Guardrails and human review

AI can draft fast and still be wrong with total confidence, so no output counts until a named person has checked it. Two habits make that practical.

First, review where it matters most. If people must approve everything, they stop reading and click yes. Set a threshold instead: every client-facing message, and any payment over a set amount. In one of my n8n practice lessons, AI reads each invoice PDF and logs it, and anything over $1,000 waits for the owner's Approve button.

Second, write your hardest rule into the agent in plain words, with a next step, and test the risky question before anyone else sees it.

7. Decide your incident plan now

It will happen, so decide ahead of time. Mine has five steps:

1. Stop and tell. The person tells the policy owner the same day. Nothing is deleted first, so you know what happened.
2. Record it. Which tool and account, what was pasted, when, and whether a shared link or connected app was involved.
3. Contain it. Delete the chat and any shared link, check training and memory settings, and remove any unapproved tool from the device.
4. Decide who else to tell. Check your client contract. If personal data was involved, ask a qualified adviser whether you must notify anyone.
5. Fix the cause. Update the cheat sheet, tools list or policy so it does not happen the same way twice.

Keep it alive

Send the policy with a read-by date, run the quiz in a team meeting, and review the tools list, settings and incidents every six months. AI tools change fast, and a policy that names last year's tools protects nobody.

Get the free kit

The page has three free downloads: a data classification cheat sheet, an approved AI tools list you can open in Google Sheets or Excel, and a one-page AI policy template with ten sections. Replace everything in the brackets and you have a first draft the same day. The free course, AI Governance for Small Business, takes a little over an hour, and you practice each piece in a safe copy of the real tool.

Read the full starter kit and grab the templates: AI Governance for Small Business

Now the client asks a fair question: "What do you keep in AI tools?" The partner wants it cleaned up before Friday.

This is the situation I built a free practice lesson around, and the surprise in it is simple. The details are in two places, not one. There is the chat, and there is ChatGPT's memory. Deleting the chat does not delete the memory. And archiving the chat deletes nothing at all.

This is a cleanup guide, not a settings checklist. It is what to do after the paste has already happened. The steps match ChatGPT on a Plus account as I checked it in October 2026. Screens change, so check your own account too.

Step 1: Find the chat and delete just that one

Your chats are listed in the left sidebar under Recents, newest first. ChatGPT picks the names, so look for one that sounds like the client work. The magnifying glass at the top of the sidebar searches them.

Open the chat and check it is the right one. Then click the three dots at the top right. The menu has Rename, Pin, Archive, Delete and a few more. Click Delete. ChatGPT warns that this will permanently delete the chat and that it can't be undone. Click Delete chat.

That removes one conversation and leaves the rest of your history alone.

Step 2: Archive what you keep, delete what you should not hold

Archive and delete look similar and do very different things. Archive only moves a chat out of the sidebar. There is no warning because nothing is lost. The chat is still stored in your account.

My rule for a firm is short: archive what you keep, delete client details you should not be holding. A month-end checklist chat with no client data is a good archive. The chat with the account number is a delete.

To find archived chats, click your name at the bottom left, then Settings, then Archived chats. Each one has Unarchive and Delete next to it.

Step 3: Clear the memory, one memory at a time

This is the step most people miss. In Settings, open Personalization. Under ChatGPT memory you will see the Enable memory switch, which controls whether ChatGPT uses your chats, files and connected apps to personalize answers. Next to it is a link to your saved memories.

Open the saved memories. In my practice account there were three. Two were harmless. One named the client and the balance. Point at that one, click its three dots and choose Delete. Only that memory goes.

There is also a Delete all memories option in the menu at the top of the list. Do not use it for a cleanup. It wipes the useful memories too, and you are trying to remove one problem, not start over.

If a login is shared by more than one person, many firms simply turn Enable memory off. One person's client note should not be able to show up in someone else's chat.

Step 4: Know where "Delete all chats" lives, and leave it alone

In Settings, Data controls, there is a Delete all chats button. Know where it is and do not press it. It removes every chat at once, including the ones your firm needs for its records, and it cannot be undone. A cleanup is one chat at a time.

Before any cleanup bigger than a chat or two, use Export data on the same page. It emails you a file of your chats, so the firm keeps a record of what existed.

While you are there, check Improve the model for everyone and switch it off on any account used for client work.

Step 5: Use Temporary Chat for the next sensitive job

On a new chat, the icon at the top right turns on Temporary Chat. The screen changes to "Temporary chat" and tells you the chat will ignore memory, plugins and custom instructions, and won't appear in your history.

That means nothing to clean up afterward in your sidebar or in memory. But read the line at the bottom of the screen: ChatGPT says it may keep a copy for up to 30 days for safety. So Temporary Chat is safer, not secret. The habit that really protects you is still the old one: replace names and account numbers with placeholders like [CLIENT] before you paste.

What this is worth

Ten minutes of cleanup replaces an awkward "we're not sure" when a client, an insurer or an auditor asks what your firm keeps in AI tools. It also gives you a repeatable answer: here is where details live (chats and memory), here is how we remove them, and here is what we do differently next time.

I would write it on a one-page checklist and pin it where people work:

1. Delete the chat that has the client details.
2. Archive the chats you keep.
3. Delete the one saved memory, not all of them.
4. Export before anything big, and never use Delete all chats.
5. Use Temporary Chat for sensitive one-offs, with placeholders instead of real names.

Practice it before you need it

Reading steps is a start. Doing them is how they stick. The free lesson walks you through all of this in a practice copy of ChatGPT, so nothing touches your real account. It takes about ten minutes and 26 steps, and it ends with a cleanup checklist you can copy.

If you want the settings side first, I wrote a separate checklist on ChatGPT data controls for a team. This article is the other half: what to do once a paste has already happened.

Try the free lesson: Clear Client Data Out of ChatGPT Before It Becomes a Problem