The role of the minute taker can be traced back to the scribes of the early Mesopotamian or Egyptian civilizations. As administrative systems and writing technology have evolved, so too has minute-taking -- through the documenting of early church and legal proceedings in the Middle Ages, the streamlining of writing with the invention of the typewriter in the 19th century, or the rise of the computer in the 20th century. The most recent step in this long evolution is the rise of artificial intelligence systems, that can listen to complex discussions and automatically draft detailed, accurate meeting minutes.
In business, minute-taking can be critical in meetings, to record discussions, decisions, and next steps. In some cases, such as board meetings, minutes can be a legal or regulatory requirement. The minute-taker may be a nominated participant or even a formal function within the organization, such as a board assistant or company secretary. Minute-takers can also be contracted from third parties, such as a Corporate Service Provider (CSP).
In this article, we cover the role of the minute-taker and discuss
- how you can use AI to automatically draft minutes
- and how CSPs already use AI meeting-minute software.
Why is minute-taking important?
Regulatory compliance: In many jurisdictions, minutes are required for formal gatherings, such as board meetings, where minutes document that the board is fulfilling oversight responsibilities.
Project management: in project management, meeting minutes record tasks, responsibilities, and deadlines. This helps to avoid or resolve disputes and keeps projects on track and within budget.
Accountability: Documenting action items and assignees in meetings generally helps to hold participants accountable for their commitments.
Historical record and transparency: Meeting minutes promote transparency in an organization and serve as a historical reference of past decisions and the rationale behind them.
The role of the minute-taker
Effective minute-takers usually perform the following functions:
- Record key points: Document the main discussion points, decisions, and action items. Instead of transcribing an entire conversation, minute-takers should condense lengthy discussions into concise, relevant notes.
- Identify contributions and commitments: Where relevant, note who made key points and who has been assigned tasks, to ensure accountability.
- Maintain a repository of past minutes: Make past meeting minutes accessible for planning and as a reference for on past decisions.
- Follow up: Promptly distribute minutes to relevant stakeholders after a meeting.
What makes for a good minute-taker?
Key competencies that distinguish a good minute-taker are:
- Good summarization: A good minute-taker skilfully summarizes key points of discussions and outcomes. Identifying the right level of detail can be a difficult, nuanced task that requires experience and subject-matter knowledge.
- Objective, clear writing: Minutes should be an unbiased record of a meeting. The minute-taker should document discussions without adding personal commentary. Ambiguous writing that leaves room for misinterpretation can lead to disputes.
- Clear presentation: Using a well-structured, standardized template across meetings makes it easy to quickly understand meeting minutes.
- Discretion with sensitive information: Handle sensitive information with care and confidentiality. Minute-takers should be mindful of who can access minutes and ensure sensitive information is protected.
How AI is changing the role of the minute-taker
Documenting a complex discussion in real-time and in detail can be a difficult task since the minute-taker must listen, write, and sometimes participate in the discussion, at the same time. Recent advances in artificial intelligence solve that problem: neural networks trained on text and audio with Deep Learning can be used to create software that transcribes meetings with high accuracy, and then automatically converts a verbatim transcript into professional meeting minutes.
Minute-takers are fast to adopt this new technology. They use AI software to transcribe meetings and create a first draft of the minutes. Before meetings, minute-takers may provide the AI with a meeting agenda to help structure the meeting minutes. During the meeting, the minute-taker only takes selective notes, to highlight which points may need particular attention in the meeting minutes. After the meeting, the minute-taker then reviews the AI’s draft minutes for completeness and accuracy, and then shares the minutes with stakeholders for final approval. The role of the minute-taker is thus fast evolving into one of human-AI collaboration.
What to look for in AI meeting-minute software
- Accuracy
- The accuracy of a minute-taking system is controlled by two factors: the transcription accuracy, and the accuracy of converting a verbatim transcript into draft minutes.
- Errors in the transcript will carry over into errors in the automatic draft minutes minutes. Common issues are mistranscription of jargon, acronyms, names, or other proper nouns in business language.
- Ensure that the software uses a capable AI to convert the transcript into draft minutes. The AI needs to be able to distinguish what is important from what can be left out for brevity. That typically requires a Large Language Model, sometimes also known as World Model. Such AI models go through extensive training to attain domain knowledge on all common subjects.
- Speaker recognition
- Ensure that the system accurately distinguishes between speakers, to avoid misattributions. The quality of this speaker diarization capability varies widely across systems.
- Detailed Minutes
- It is better to use a system that drafts minutes that are too detailed, than a system that summarizes too much, since it is easier for a human reviewer to redact superfluous detail, than it is to check minutes that lack detail for completeness.
- Support for different meeting settings
- Often, meetings occur in different settings: they can be virtual, using different platforms (eg. Microsoft Teams, Zoom, or Google Meet) or face-to-face, for instance in a conference room.
- They can also be hybrid, with physical meeting rooms joining virtual meetings. Teams’ or Zoom’s built-in transcription does not work with hybrid or in-person settings.
- A stand-alone system that works across virtual platforms, hybrid settings, and in in-person meetings is most versatile. For in-person meetings, a vendor that provides the option of a mobile application is best.
- Human review
- Even with the best systems, automatic transcripts and meeting minutes can contain errors – so it is important to always review automatic drafts before approving and sharing final minutes.
- For software, that means that no automatic sharing of meeting minutes should take place so that the minute-taker has the opportunity to review.
- Speed
- Minutes should be automatically drafted live in the meeting or immediately after the call. The speed with which AI can operate is one of the key benefits of an automated system. Do not use a tool that bottlenecks that time-to-delivery with slow software.
- Privacy and security
- Meeting minutes can contain highly sensitive and confidential information. Use software that is secure.
- Look for: minimal data processing – ideally, no storage of audio or video recordings (biometric data) – and custom data hosting options for control (geography, private cloud or on-premise storage, custom data retention timeline).
- Carefully review privacy policies and terms of use. Ensure that data is not used to train artificial intelligence systems.
- Software tools should be regularly audited by a third party and be subject to penetration testing.
One tool that covers these points is Cogram – a software that specializes in automatic minutes for in-person (face-to-face), hybrid, and virtual meetings. The Cogram platform integrates with Zoom, Microsoft Teams, Google Meet, calendars, and email, and includes mobile applications for Android and iOS phones. Beyond meeting minutes, Cogram can draft project reports based on multiple meetings and email threads. The system is multi-lingual.
Cogram emphasizes privacy and security: the company offers different data hosting options, including private cloud storage, and custom retention timelines. The software does not store audio or video recordings and can be configured to automatically delete transcripts after minutes are drafted – so that no verbatim machine-generated record of a meeting is stored.
How corporate service providers are adopting AI meeting-minute software
Corporate Service Providers (CSPs) offer a range of professional services to businesses, particularly those related to corporate governance, compliance, administration, and operational management. CSPs are often used by companies to handle administrative and regulatory tasks that they may not have the expertise, time, or resources to manage internally. Those services often include minuting official meetings, such as board meetings.
CSP firms are fast adopting AI for meeting minutes, to handle a higher volume of meetings without compromising on accuracy or time-to-delivery for their clients. Typically, professional minute-takers such as board assistants use AI to rapidly create a first draft of meetingng minutes, which they then review before sharing finalized minutes with their client.
Wrapping up
The evolving role of the minute-taker is representative of the ongoing trend towards human-AI collaboration – one that significantly increases productivity, enables firms to serve a larger client base, and deliver more value to the customer.
