Your best AI data is the meeting nobody recorded

4:51

Should your nonprofit record meetings for AI?

Yes, for most meetings. Every conversation your team has holds decisions, donor details, program stories and promises. When nobody records it, AI can't use any of it, and neither can you two weeks later. Recording turns talk you already do into context AI can work from.

Colleen and I record nearly everything. On a September recording she called it the most important AI habit there is:

recording your meetings and using those conversations as data is the number one skill that you need to learn

Colleen Cook · Recorded conversation, Sep 8, 2026

In October she put our number at about 95% of our meetings, conversations and client work. The few we skip are the ones that shouldn't be recorded at all, and I'll get to those.

Why do recorded conversations matter so much for AI?

AI is only as good as the context you give it, and talk is where most of your context lives. Your strategy, the name of a donor's daughter, the reason the board passed on an idea last spring. Most of it is said out loud and never written down, so your AI never sees it.

I got onto this because of how much I was losing:

We are always talking. We just don't do a good job of retention. And so that got me thinking like, Oh man, I wonder if we could just retain all conversations all the time.

Scott Williams · Recorded conversation, Sep 10, 2026

On our agency team, I'd say notes and follow-up were half of some people's jobs, and we had several people doing that work. Now the recording does the remembering.

There's a second benefit nobody expects. When you know it's recording, you stop typing and start listening. I used to sit about ten seconds behind every conversation, trying to capture what was just said. Now I just talk. Colleen says being able to concentrate fully is the most valuable thing recording gives her.

How do you ask for consent to record a meeting?

Ask at the start, in plain words, and say what the recording is for. Most people say yes when they hear it helps you follow through on what you promise them. If someone says no, respect it right away and capture the meeting another way.

Here's how I said it in September:

Then I explain why. It helps us do good work for them afterward. I don't believe anyone has ever told me no. Colleen guesses it happens to her about once in a hundred conversations, and she has a simple fallback:

what if they don't want you to record? No problem. Don't record that and get in your car and debrief it on your drive back.

Colleen Cook · Recorded conversation, Sep 8, 2026

Two more consent basics. The first is the law. Recording rules vary by state. Ohio, where we work, is a one-party consent state, and some states require everyone in the conversation to agree. Colleen's advice is to always defer to what your state says, and when in doubt, ask everyone. For anything sensitive, ask your attorney.

The second is the moral question, which matters more. Are you treating people well? Asking first, every time, is how you answer yes.

Which meetings should you record, and which should you skip?

Record the meetings where decisions, ideas and promises happen. Skip the ones where someone's private information or a sensitive moment is the whole point. Colleen's rule is that we record every conversation unless it's something we really shouldn't. That one sentence covers most of the gray area.

Here's how we sort it:

Where should you store meeting recordings and transcripts?

Keep them in one place your team controls, on paid plans, with access limited to the people who need it. Check the data settings on every tool in the chain so your transcripts aren't used to train a company's model. Then point your AI at that one home.

Colleen explained the privacy piece simply. Free AI tools can use what you put in them. Paying is how you get out of that agreement, and paid tiers give you the option to opt out of training. When she said it in September, Team and Enterprise plans had it off by default. Plans change, so check yours.

Our storage basics:

• One home. Our transcripts flow into our organization's AI brain in Notion. Colleen likes it because it's permission controlled and works smoothly with Claude.
• Paid plans everywhere. The note taker, the AI, the storage. Training on your data switched off.
• Access by role. Development sees donor conversations. Not everyone needs everything.
• The full transcript. I don't care much about the auto-summary. I was there. The transcript is what AI needs.
• A written line. Put what you never record into your first AI policy so nobody has to guess.

What do you do with meeting transcripts once you have them?

Give them jobs. A transcript can become a follow-up email, a task list, a campaign plan or a month of posts. The more often your AI reads your transcripts, the less your team has to hold in their heads.

The simplest job is catching promises:

Other jobs we run every week:

1. Nightly filing. Every night a process grabs our transcripts and sorts what matters into the right part of our brain.
2. No retyping. When Colleen and I decide something on a call, she tells AI to go pull the conversation and make the change.
3. Campaign plans. Colleen recorded herself talking through what she wanted a year-end campaign to be. That transcript started the whole campaign.
4. Content. This fall I ended a transcription with Colleen, opened Claude and told it we'd just chatted and needed videos. By Oct 13 we had made over 300 pieces of content from about 13 recorded conversations.
5. Picking up the thread. Colleen can come back to a conversation months later as if it just happened.

How do you start recording meetings this week?

Start with one recurring meeting and one tool. Turn on an AI note taker, ask for consent at the top, and afterward paste the transcript into your AI with one request. List every decision, owner and promise. Do that for two weeks before you add more meetings.

Here's a prompt to copy:

``` Here is the transcript of our [meeting name] from [date]. List every decision we made, every task with its owner, and every promise anyone made to someone outside our team. Flag any commitment that has no date. Then draft a short follow-up email to the attendees for me to review before it goes out. ```

Once the habit sticks, transcripts fuel the five workflows we'd automate first and fit right into your first 30 days with AI. Recorded donor conversations also make a thank-you system far more personal. You'll find more like this in our implementation articles.

If you want what we're learning as we go, join the newsletter. And if you'd like to build this habit alongside other nonprofit teams, with Colleen and me in the room, get on the cohort waitlist.