Meeting Notes AI Prompts: 5 Templates for PRDs, Issues, and Emails
Still replaying the meeting recording, cleaning up your notes, pasting them into ChatGPT, and then copying the answer into your PRD and email templates one piece at a time? When you paste in a raw transcript and just say "summarize this," the output changes every time, and fixing it can take as long as the meeting itself.
A meeting notes prompt is an AI instruction that gives the model your transcript as context and also spells out the format, audience, and length of what you want back. This guide shares five templates that PMs, planners, and marketers can use right after a meeting to draft a PRD, an issue list, or a client follow-up email. It also shows how to run the same templates on top of speaker-separated transcripts with daglo and Board Chat.
What makes a meeting notes prompt work?
The quality of a meeting notes prompt depends mostly on how organized the input is. If you first separate who said what (speakers), what was agreed (decisions), and what is still open (open questions), the AI has far less to guess.
- Speakers: separating client requests from internal opinions keeps the wrong content out of your email and your spec.
- Decisions: only confirmed items should make it into the body of a PRD.
- Open questions: anything undecided belongs in the issue list or the next meeting's agenda.
That is why Template 1 below is a prep prompt that sorts the transcript into these three buckets before you create anything else. The other four templates use that sorted output as their input.
5 meeting notes prompt templates by situation
The table below lists the five follow-up documents people create most often after a meeting, along with the core sentence of each prompt. Swap out the bracketed parts and they work in any AI chatbot.
| Situation | Output | Core prompt |
|---|---|---|
| 1. Right after the meeting | Decisions, open questions, to-dos | Using each speaker's statements as evidence, sort the transcript below into decisions, open questions, and to-dos. |
| 2. Feature discussion | PRD draft | Using only the decisions, draft a PRD for [feature] with background, goals, requirements, out of scope, and success metrics. |
| 3. Sprint or QA meeting | Issue list | Turn the open questions and reported problems into a table with title, description, priority, and raised by. |
| 4. Client meeting | Follow-up email | Write a follow-up email to [client contact]: what we agreed, what we will do next, and questions for them to confirm. |
| 5. Next meeting prep | Agenda | Turn the open questions into agenda items and add one decision question under each. |
End every template with "Don't include anything that isn't in the transcript." It noticeably cuts down on plausible-sounding details the AI makes up.
Example prompts for a PRD and an issue list
The PRD and issue list prompts pull different parts out of the same transcript. The key is to scope them clearly: the PRD gets only decisions, while the issue list gets only open questions and reported problems.
You are the PM for [product]. Using only the "Decisions" section of the meeting summary below, write a PRD draft for [feature]. Use this order: background, goals, user scenarios, requirements, out of scope, success metrics. If there isn't enough evidence for a section, leave it empty and write "Not decided in the meeting."
For issue lists, specifying a table format makes it easy to move the output into Jira, Linear, or whatever tracker your team uses. Add these three lines for more consistent results:
- Priority rules: something like "High if it affects the launch date" keeps the AI's judgment consistent.
- Raised by: knowing who flagged each problem speeds up follow-up.
- Merge duplicates: add "If the same problem was described in different ways, merge it into one item."
A client email prompt and a review checklist
A client follow-up email is a record of what both sides agreed to, so reviewing the draft matters most. Setting the recipient, purpose, tone, and length in the prompt cuts down how much you need to fix.
- Numbers and dates: double-check quotes, timelines, and quantities against the transcript.
- Scope of commitments: make sure "we'll look into it" didn't quietly become "we'll do it."
- Internal comments: confirm that nothing said after the client left the call ended up in the email.
Most internal comments stay out of the email automatically when the transcript you feed in already separates speakers.
Record with daglo and ask Board Chat the same templates
When you record a meeting with daglo, you get a speaker-separated transcript and a summary, and Board Chat lets you ask the templates above with that recording as context. You no longer need to copy the transcript into a separate AI window.
- Start recording in the daglo app when the meeting begins, or upload the recording of a video call.
- When the transcript is ready, check the speaker names and skim the summary.
- Ask Board Chat Template 1 to sort out decisions, open questions, and to-dos.
- Then ask the templates you need next: PRD, issue list, or email.
Board Chat answers from what is in the recording. If an answer feels vague, follow up with something like "Who said that?" to check the original transcript. Keep your favorite templates in a notes app and paste them in after every meeting.
#meetingnotes #AIprompts #prompttemplates #PRD #productmanagement #AImeetingnotes #BoardChat #daglo
Where talk becomes work.
Get started with 1,000 credits a month — no credit card required.