AI for finance meetings and board prep

AI for finance meetings and board prep means using AI note-takers and summary tools for MT meetings, audit calls, and board meetings — recording, summary, and action list automatically, regardless of speaker. For CFOs and senior controllers spending six to ten hours a week in finance meetings this yields a structural two to four hours of time saved per week — provided consent and privacy choices are sorted upfront.

The most underestimated AI use for CFOs, controllers, and finance managers isn't a brilliant prompt or a complex agent. It's an AI note-taker that automatically records, summarizes, and converts every MT meeting, audit call, and board meeting into an action list. For anyone spending six to ten hours a week in finance meetings — not unusual for a CFO or senior controller — that's a structural two to four hours of time saved per week, with no extra effort.

For finance context there's one extra dimension that weighs less elsewhere: consent and privacy. An AI listening in on an audit meeting, an M&A call, or a conversation about headcount planning has a different risk profile than a marketing standup.

The landscape

Three tool categories, each with a different starting point.

1. Dedicated AI note-takers — Fireflies, Otter, Read.ai, Fathom

Specialized services that dial in or hook into your calendar and automatically record every Zoom, Teams, or Meet call. Output: transcript (with speaker identification), summary, action list, soundbites, a searchable archive. API access for further processing.

Strong for finance: full specialization, best transcript quality (especially on Dutch-English mixed calls with external auditors or investors), integrations with Slack, Notion, mail. Weak: extra tool and subscription, external processor (privacy consideration), governance question on who can see what — particularly relevant for finance.

2. Built into the meeting software — Teams Intelligent Recap, Zoom AI Companion

The video platforms deliver a note-taker themselves. No extra tool needed, no extra bot in the call.

Strong for finance: no extra vendor, lower privacy bar (same DPA as the call itself), automatic link to your calendar. For MS365 finance teams often the first pick. Weak: less sharp than specialized tools on specific finance terminology, more limited integrations, often only in higher license tiers.

3. After-the-fact processors — Claude, Copilot, ChatGPT with a recording

You record yourself (or download the transcript) and feed it to a general model with a good prompt or skill. Full control over processing and output.

Strong: maximum control, can run in your own enterprise environment, flexible to any desired format. Weak: manual, no real-time integration, requires having a recording or transcript already. For occasional processing of a specific board call: usable.

The real win — from minutes to follow-up

Transcripts are barely useful for finance. Summaries are reasonably useful. Follow-ups are very useful. Every modern note-taker, with the right workflow skill behind it, can deliver:

• Action list per finance role — "For the controller: work out variance commentary for category X; deadline quarter close." Direct to push into Notion, Asana, or mail.
• Draft board update — to the supervisory board or investors, with summary + action items + open questions.
• Agenda for the next MT — "Open points from last meeting" pre-filled.
• Audit-meeting trail — for calls with the external auditor: a structured record that fits your audit file.
• Searchable finance-meeting archive — "When did we decide to take provision X?" answered in 30 seconds.

The tool delivers the first layer. The value sits in the workflow around it: a skill that takes transcript + attendee list + finance context (client, project, quarter) and builds usable output.

Consent and transparency — sharper for finance

Recording an AI note-taker without participants' consent is legally dubious and relationally damaging. For finance meetings extra:

• Always announce in the calendar invite and at the start of the meeting. Many tools do this automatically via a visible bot or an opening message.
• Option to opt out. Participants must be able to say "don't record today" without fuss. With audit meetings with the external auditor: align in advance on whether recording is allowed. Often the answer is yes, provided transparent — sometimes no, then respect.
• Clarity on what happens — who gets access to the transcript, how long is it kept, who gets the summary.

For fundamentally sensitive conversations — a salary discussion, an M&A call with an external party, a legal session on a dispute — usually no AI note-taker, full stop. The relational and legal risks don't justify the time gain.

Privacy and GDPR in finance context

Every AI note-taker is a processor under GDPR. For finance:

• Data processing agreement (DPA) required. Fireflies, Otter, Read.ai have standard DPAs — read them. Microsoft Teams and Zoom have their own DPAs as part of the main contract.
• Data residency — where do the servers sit, where do the transcripts live? For EU clients you want EU hosting or an explicit guarantee that finance-meeting data doesn't go to the US.
• Retention periods — set transcripts to delete after 30-90 days, unless explicitly archived. Audit-meeting transcripts keep longer (matching your internal-control retention); MT meetings shorter.
• Training on your data — free or consumer tiers may use your transcripts for product improvement. Business tiers should exclude this; check this explicitly in your contract.
• Access control — who in finance can see transcripts of which meetings? Especially relevant for meetings discussing personal data of staff or numbers of clients.

For finance meetings with special categories of personal data (health, legal dispute, salary level) — no AI note-taker, or only in a fully on-premise setup.

Practical setup for finance staff with many meetings

For a CFO or senior controller with 8-12 meetings a week:

1. Pick one note-taker and stick to it. Switching tools makes your archive unusable.
2. Link to your calendar, not manually per call. Teams Intelligent Recap or Fireflies can do this via your Outlook calendar.
3. Set up one standardized output — "transcript in OneDrive folder, summary in Notion, action list in Outlook tasks, mail in drafts." Via Power Automate, n8n, or an MCP workflow.
4. Write one skill — "turn transcript X into a board update in my style, referencing quarter Y." Refine after each use.
5. Review after two weeks. Which tool misses something you find important? Which output do you not use? Prune what doesn't work.

Often overlooked — meetings you don't attend

An underestimated finance use case: colleagues or partners have a meeting relevant to you, but you don't have to be in it — a commercial MT, an operations meeting with budget impact, a customer call influencing the forecast. A shared Fireflies archive or Teams Recap lets you read the transcript in five minutes — or have the AI extract the finance-relevant highlights for your context.

This in practice halves the number of meetings where "it would be handy if finance were there." Delegate the meeting to someone with decision authority, then read the finance summary. Saves a CFO hours per week; does require the organization to accept this way of working.

Where the hype line sits

AI note-takers are no substitute for attention in the meeting itself. The presence of a bot changes the dynamic of a finance conversation, especially on sensitive topics — people get more careful, less spontaneous. For some conversations that's no problem; for others it is. With strategy brainstorms, performance reviews, or coaching with a team member: note-taker off, conversation open.

And: a summary is a summary. The model interprets, filters, can miss nuance. Fine for most finance meetings; for decision conversations where every phrasing counts (formalizing audit findings, confirming an agreement with an external auditor) review the transcript yourself before storing it.