# Meeting Notes: Q3 Product Roadmap Strategy ## Meeting Summary Sarah, Mike, and Jen met in person for approximately 45 minutes to align on Q3 product roadmap priorities. The team debated a mobile-first approach versus focusing on the enterprise dashboard redesign, ultimately agreeing to prioritize the dashboard work. Next steps include scoping the dashboard effort and coordinating with the design team. ## Attendees - Sarah - Mike - Jen (note-taker) ## Key Discussion Points **Q3 Product Roadmap Direction** - Reviewed proposed priorities for the upcoming quarter. **Mobile-First Approach** - Mike raised concerns about pursuing a mobile-first approach, citing limited engineering bandwidth. - Trade-offs between mobile investment and other priorities were weighed. **Enterprise Dashboard Redesign** - Sarah advocated for prioritizing the enterprise dashboard redesign ahead of other initiatives. - Team aligned that dashboard work offers more immediate value given current capacity constraints. ## Decisions Made - Deprioritize the mobile-first initiative for Q3 due to engineering bandwidth limitations. - Prioritize the enterprise dashboard redesign as the leading Q3 workstream. ## Action Items | Owner | Task | Due Date | |-------|------|----------| | Mike | Scope the enterprise dashboard redesign work | Next Friday [unclear — specific date not provided] | | Jen | Set up a follow-up meeting with the design team | Not specified | ## Open Questions / Follow-ups - What is the exact calendar date for Mike's scoping deliverable? [unclear] - When will the follow-up with the design team be scheduled, and who should attend? - Will mobile-first be revisited in Q4, or removed from the roadmap entirely? - Are there specific engineering bandwidth figures or constraints to document for future planning?
The Best AI Note Takers for In-Person Meetings
Tested prompts for ai note taker for in person meetings compared across 5 leading AI models.
BEST BY JUDGE SCORE
Claude Opus 4.7
8/10
You're in back-to-back client meetings, a team standup, or a field interview, and you need accurate notes without burying your face in a laptop. The challenge with in-person meetings is that no bot automatically joins a physical room the way it joins a Zoom call. You need a different workflow: record the audio yourself, then hand it to an AI to transcribe and summarize.
This page shows you exactly how to do that. We tested a specific prompt across four leading AI models to see which one produces the cleanest, most actionable meeting notes from a raw audio transcript or rough spoken summary. The comparison table below shows the differences in structure, completeness, and follow-up clarity.
Whether you're a consultant capturing client requirements, a manager running a weekly team sync, or a researcher conducting interviews, the right AI note-taking setup turns a voice recording into a structured document in under two minutes. Here's what works, what doesn't, and which model to trust with your real meetings.
When to use this
This approach fits any situation where a meeting happens in a physical space and you can record audio on a phone or dedicated device. It works best when the meeting has clear decision points, action items, or follow-ups that need to be tracked. If you can hit record before anyone starts talking, this workflow delivers structured notes faster than manual writing.
- Client discovery or requirements meetings at an office or job site
- Team standups, retrospectives, or planning sessions in a conference room
- Field interviews, user research sessions, or focus groups
- One-on-one performance or coaching conversations you need to document
- Sales calls, vendor negotiations, or partnership discussions held in person
When this format breaks down
- You cannot legally or ethically record the conversation without all-party consent in your jurisdiction. In many U.S. states and most of the EU, recording without consent is illegal.
- The audio quality is too poor to transcribe accurately, such as in loud restaurants, open-plan floors with heavy background noise, or outdoor environments with wind interference.
- The meeting is highly confidential and your organization prohibits uploading audio or transcripts to third-party AI services due to data security or NDA obligations.
- You need real-time live notes during the meeting itself. This workflow is post-meeting: record first, process after.
The prompt we tested
You are an expert meeting notes assistant specialized in capturing and structuring notes from in-person meetings where audio may be imperfect, multiple speakers overlap, and context matters. Follow these rules strictly: Output structured notes in Markdown with these sections: **Meeting Summary** (2-3 sentences), **Attendees**, **Key Discussion Points** (bulleted, grouped by topic), **Decisions Made**, **Action Items** (table with Owner, Task, Due Date), and **Open Questions / Follow-ups**. Keep language concise and professional, infer reasonable context when audio is unclear but flag uncertain items with [unclear], and never invent names, dates, or commitments not present in the input. Here is the raw meeting content (transcript, recorded audio notes, or rough jotted notes from an in-person meeting): Rough notes from today's in-person strategy meeting in the conference room: Sarah, Mike, and I (Jen) met for about 45 min to discuss the Q3 product roadmap. Mike pushed back on the mobile-first approach citing engineering bandwidth, Sarah agreed we should prioritize the enterprise dashboard redesign first, and we decided Mike will scope the dashboard work by next Friday while I set up a follow-up with the design team. Produce polished, well-organized meeting notes now.
Model comparison
| Model | Quality | Cost | Latency | Tokens |
|---|---|---|---|---|
| Claude Opus 4.7 | 8 | $0.0586 | 9.5s | 456 / 690 |
| Claude Haiku 4.5 | 7 | $0.0022 | 4.0s | 290 / 378 |
| GPT-5.4 | 8 | $0.0049 | 4.6s | 263 / 281 |
| Gemini 2.5 Pro | 8 | $0.0019 | 9.8s | 264 / 308 |
| Grok 4.1 Fast Reasoning | 7 | $0.0002 | 6.0s | 408 / 215 |
Outputs
# Meeting Notes: Q3 Product Roadmap Strategy **Meeting Summary** The team met to discuss Q3 product roadmap priorities and encountered a resource constraint conflict regarding the mobile-first approach. After discussion, the group agreed to prioritize the enterprise dashboard redesign first, with Mike to scope the work and Jen to coordinate with the design team. **Attendees** - Sarah - Mike - Jen **Key Discussion Points** **Product Roadmap Priorities** - Mobile-first approach was proposed but encountered resistance - Mike raised concern about engineering bandwidth constraints for mobile-first initiative - Sarah advocated for prioritizing the enterprise dashboard redesign instead - Team consensus: enterprise dashboard redesign should be the Q3 priority **Resource & Scope Considerations** - Engineering bandwidth is a limiting factor for parallel initiatives - Dashboard redesign identified as more feasible within current constraints --- ## Decisions Made - **Priority Decision:** Enterprise dashboard redesign will be prioritized for Q3 over mobile-first approach - **Approach:** Address dashboard work first, then reassess mobile-first feasibility --- ## Action Items | Owner | Task | Due Date | |-------|------|----------| | Mike | Scope the enterprise dashboard redesign work | Next Friday | | Jen | Set up follow-up meeting with design team | [unclear—assume early next week pending Mike's scope] | --- ## Open Questions / Follow-ups - Timeline for mobile-first approach—will this be deferred to Q4 or reconsidered once dashboard is scoped? - What is the specific scope of the enterprise dashboard redesign? - When should the design team follow-up meeting occur?
# Meeting Summary The team met for approximately 45 minutes to discuss the Q3 product roadmap. A mobile-first approach was challenged due to engineering bandwidth constraints, and the group aligned on prioritizing the enterprise dashboard redesign first. # Attendees - Sarah - Mike - Jen # Key Discussion Points ## Q3 Product Roadmap - Reviewed priorities for the Q3 product roadmap. - Discussed whether to pursue a mobile-first approach or focus on other initiatives first. ## Resourcing and Feasibility - Mike raised concerns about the mobile-first approach, citing engineering bandwidth limitations. - Sarah supported shifting priority to the enterprise dashboard redesign. ## Priority Alignment - The group aligned on prioritizing the enterprise dashboard redesign ahead of the mobile-first effort. # Decisions Made - The enterprise dashboard redesign will be prioritized first in the Q3 roadmap. - The mobile-first approach will not be the immediate priority due to engineering bandwidth concerns. # Action Items | Owner | Task | Due Date | |---|---|---| | Mike | Scope the enterprise dashboard work | Next Friday | | Jen | Set up a follow-up meeting with the design team | [unclear] | # Open Questions / Follow-ups - What is the timeline for the follow-up with the design team? [unclear] - When should the mobile-first approach be reconsidered within the roadmap?
Here are the structured meeting notes:
### **Meeting Notes: Q3 Product Roadmap Strategy**
**Meeting Summary**
The team convened to finalize the Q3 product roadmap priorities. After discussing a potential mobile-first approach, the group decided to prioritize the enterprise dashboard redesign due to engineering bandwidth constraints. Next steps were assigned to scope the dashboard project and engage the design team.
**Attendees**
* Sarah
* Mike
* Jen
**Key Discussion Points**
* **Q3 Roadmap Prioritization**
* A mobile-first product strategy was initially discussed for the upcoming quarter.
* Mike raised concerns that the engineering team currently lacks the bandwidth to effectively execute a mobile-first initiative.
* Sarah supported this concern and proposed prioritizing the enterprise dashboard redesign as a more achievable and high-value goal for Q3.
**Decisions Made**
* The enterprise dashboard redesign will be the primary focus for the Q3 product roadmap, deferring the mobile-first approach.
**Action Items**
| Owner | Task | Due Date |
| :--- | :--- | :--- |
| Mike | Scope the engineering work for the enterprise dashboard redesign. | Next Friday |
| Jen | Schedule a follow-up meeting with the design team to discuss the dashboard redesign. | Not specified |
**Open Questions / Follow-ups**
* A follow-up meeting with the design team is required to kick off the redesign process.
**Meeting Summary** Jen, Sarah, and Mike held a 45-minute in-person strategy meeting to discuss the Q3 product roadmap. Mike raised concerns about engineering bandwidth for a mobile-first approach, leading to agreement on prioritizing the enterprise dashboard redesign. The team assigned scoping tasks to Mike and a design team follow-up to Jen. **Attendees** - Jen - Sarah - Mike **Key Discussion Points** - **Q3 Product Roadmap** - Debate on mobile-first approach; Mike cited limited engineering bandwidth as a blocker. - Consensus to prioritize enterprise dashboard redesign first, with Sarah's agreement. **Decisions Made** - Shift priority from mobile-first to enterprise dashboard redesign for Q3 roadmap. **Action Items** | Owner | Task | Due Date | |-------|-------------------------------|-------------| | Mike | Scope the dashboard work | Next Friday | | Jen | Set up follow-up with design team | | **Open Questions / Follow-ups** - None noted.
What makes these work
-
01Record first, then prompt
Use your phone's native voice memo app or a dedicated recorder like Otter.ai or Plaud to capture audio. Once you have a transcript or rough spoken summary, paste it directly into your AI prompt. The cleaner your input, the more precise the output. Even a rough spoken recap into your phone on the walk back to your car gives the model enough to work with.
-
02Tell the model its output format upfront
Specify exactly what you need: summary, action items with owners, decisions made, and open questions. If you leave the format open, different models will produce wildly different structures. A consistent prompt template you reuse across every meeting saves editing time and makes your notes scannable by anyone on your team.
-
03Label speakers in your transcript
If your recording includes multiple people, identify speakers before pasting the transcript: 'Client: ...', 'Me: ...', 'IT Director: ...'. Models attribute action items and decisions far more accurately when they know who said what. This is especially important for sales and client meetings where accountability tracking matters.
-
04Ask for a one-line TL;DR at the top
Add 'Start with a one-sentence meeting summary' to your prompt. That single sentence becomes the subject line of your follow-up email or the header of your notes document. It forces the model to distill the meeting's core outcome, which also helps you catch cases where the model misunderstood the context.
More example scenarios
#01 · Client requirements meeting for a software project
Input
Transcript from a 45-minute discovery call with a retail client. They want a new inventory dashboard. Key points: they need real-time stock alerts when items fall below 20 units, integration with Shopify, a mobile-friendly interface, and a go-live target of Q3. Budget is around $40k. Next step is for us to send a scoping doc by Friday.
Expected output
Summary: Client needs a real-time inventory dashboard integrated with Shopify. Trigger alerts at sub-20 unit threshold. Mobile-first design required. Budget: $40,000. Target go-live: Q3. Action item: Send scoping document to client by Friday. Owner: [Your name].
#02 · Weekly team standup in a startup office
Input
Standup notes, 10 minutes. Maria finished the landing page redesign, pushing to staging today. Dev team blocked on API keys from the third-party payment provider, waiting since Monday. James will follow up with the provider directly. Sprint review is Thursday at 2pm. No other blockers reported.
Expected output
Completed: Landing page redesign (Maria), moving to staging today. Blocker: API keys pending from payment provider since Monday. Owner: James to follow up directly. Upcoming: Sprint review Thursday 2:00 PM. No additional blockers.
#03 · User research interview for a healthcare app
Input
Interview with a 58-year-old retired nurse testing a medication reminder app. She found the font too small on the main screen and couldn't locate the refill button without help. She liked the color-coded schedule view and said she would use the app daily if the text size was adjustable. Session lasted 22 minutes.
Expected output
Key findings: Font size on main screen rated too small for comfortable reading. Refill button not discoverable without guidance. Positive: Color-coded schedule view resonated strongly with participant. Feature request: Adjustable text size. Recommendation: Prioritize accessibility audit for font scaling and navigation labels before next usability round.
#04 · Sales meeting with a prospective enterprise client
Input
Met with procurement lead and IT director at a logistics company. They currently use a competitor tool and are unhappy with reporting features. Our demo of the analytics dashboard got strong positive reaction. IT director raised security concerns about data residency, wants confirmation of EU data hosting. Decision timeline is end of Q2. Follow up: send security whitepaper and EU hosting confirmation.
Expected output
Opportunity: Logistics company, dissatisfied with current vendor on reporting. Strong demo reception for analytics dashboard. Blocker: IT director requires confirmation of EU data residency. Actions: Send security whitepaper and EU hosting documentation. Decision deadline: End of Q2. Deal stage: Active evaluation.
#05 · Manager one-on-one performance conversation
Input
One-on-one with a direct report. She's been hitting her sales targets for three consecutive quarters. She wants to move into a team lead role within 12 months. We agreed she'll start shadowing team lead meetings starting next month. I'll connect her with the leadership development program coordinator by end of week. Next one-on-one in four weeks.
Expected output
Employee highlights: Three consecutive quarters on target for sales goals. Career goal: Team lead role within 12 months. Agreed actions: Begin shadowing team lead meetings next month. Manager to connect employee with leadership development coordinator by end of this week. Next check-in: Four weeks.
Common mistakes to avoid
-
Skipping recording consent
Uploading a recording of a conversation where participants did not consent creates legal exposure and breaks trust if discovered. Always disclose that you are recording at the start of the meeting. Most people agree immediately when you frame it as a tool for accurate follow-up notes.
-
Pasting raw transcript without cleaning speaker labels
Auto-transcription tools often label speakers as 'Speaker 1' and 'Speaker 2.' If you paste that directly into your AI prompt, action items get attributed to the wrong people or to no one. Take 60 seconds to find-and-replace speaker labels with actual names before prompting.
-
Using a generic summarize prompt
Prompting with 'summarize this meeting' produces a paragraph of prose that is hard to act on. You need structured output: decisions, action items, owners, and deadlines as separate fields. Generic prompts waste the model's capability and force you to manually extract what you actually need.
-
Trusting the output without reviewing it
AI models occasionally hallucinate details, confuse names, or misattribute a commitment to the wrong person. Meeting notes that get forwarded to a client or added to a project tracker without review can create real confusion. Always do a 90-second read-through before sending or filing the output.
-
Waiting too long after the meeting to process notes
The longer you wait, the harder it is to catch errors in the AI output because your own memory of the conversation has faded. Process your recording within two hours of the meeting while you can still verify whether the model captured the right decisions and action items.
Related queries
Frequently asked questions
What is the best app to record in-person meetings for AI note taking?
Otter.ai and Plaud Note are strong choices because they transcribe audio in real time and let you export text directly for AI processing. If you want a simpler setup, the built-in voice memo app on iOS or Android works fine. Record the audio, transcribe it using a tool like Whisper or your AI platform's voice input, then run your note-taking prompt on the result.
Can AI take notes in real time during an in-person meeting?
Yes, but it requires a live transcription app running on a phone or laptop placed on the table. Otter.ai, Fireflies, and Notta all offer real-time in-person transcription modes. The limitation is that speaker identification is less accurate than in post-processing, and background noise can degrade transcript quality. For most people, recording and processing after the meeting produces better results.
How do I use ChatGPT or Claude to take notes from an in-person meeting?
Record the meeting audio, then generate a transcript using a transcription service or the AI platform's voice feature. Paste the transcript into your prompt along with instructions specifying the output format you need, such as summary, action items with owners, and decisions. Both ChatGPT and Claude handle meeting transcripts well when given clear formatting instructions.
Is it legal to record in-person meetings for AI transcription?
It depends on your location and the consent of participants. In the United States, some states require all-party consent while others require only one-party consent. In the EU, GDPR requires informed consent before processing personal data including voice recordings. The safest practice everywhere is to announce the recording at the start of the meeting and get verbal agreement from all participants.
How accurate is AI transcription for in-person meetings compared to video calls?
In-person transcription is generally less accurate than video call transcription because audio quality is harder to control. Factors like distance from the microphone, overlapping speech, and room acoustics all reduce accuracy. Using a dedicated recorder placed at the center of the table, or a clip-on mic, significantly improves results. Expect occasional errors and plan to review the transcript before relying on AI-generated notes.
What should I include in my AI prompt to get good meeting notes?
At minimum, specify these four elements: a one-sentence meeting summary, a list of decisions made, action items with assigned owners and deadlines, and any open questions or next steps. If the meeting had a specific type, such as a sales call or a design review, tell the model that context so it prioritizes the right information. Reusing the same prompt template across all your meetings makes outputs consistent and easy to scan.