AI Is Already in the Boardroom – What Should Boards Be Asking?

AI is one of the most talked-about topics globally, and rightly so.

A few months ago, King Charles brought together leaders from some of the world’s biggest AI companies at Dumfries House in Scotland to discuss how AI could be developed safely and for the good of humanity. The timing was significant, coming shortly after Anthropic researcher Jacob Coxon resigned amid concerns that developers were moving too quickly towards increasingly powerful AI without sufficient safeguards.

Those conversations may feel far removed from the average Boardroom, but the underlying questions are becoming increasingly relevant there too.

This year, we have received more enquiries about how we use AI within Waymaker and, importantly, what safeguards we have in place to ensure the minutes we produce remain accurate, effective and secure. We are pleased to be asked.

We are open about the fact that AI has a place in our minute-taking processes, but how it is used matters.

The question is moving from ‘are we using AI?’ to ‘how are we using it?’

One of the biggest changes we have noticed is that AI is becoming less visible in everyday working practices. It is no longer necessarily a separate tool that somebody deliberately sits down to use, and is increasingly embedded into the way people draft, research, summarise and prepare for meetings.

Board papers may be written with significant AI assistance, and reports can be condensed before they reach the Boardroom. Directors may use it to work through lengthy information or help prepare questions ahead of a meeting. AI is also being used to support transcripts, summaries and meeting records.

None of those uses are automatically a concern, but the more important question is whether the organisation understands where AI is being used and what sits behind it.

It is worth asking:

What information is being entered?

Which platform is being used?

Whether it is a personal account or one approved by the organisation?

Who is checking what comes back out?

AI can quietly become part of a process before the governance around it has caught up. For Boards, that means the conversation needs to move beyond whether AI is permitted at all. The more useful question is whether it is being used in the right place, through the right tools and with the right level of human oversight.

Which account is actually being used?

If AI is being used for Board work, the platform itself is only part of the picture. The type of account matters too.

Board papers, minutes and reports can contain commercially sensitive information, personal data and details that were never intended to leave the organisation. Uploading that material into a personal AI account should therefore be treated very differently from using an approved business or enterprise-level platform.

At Waymaker, this is one of the boundaries we have become increasingly conscious of through our own training and use of AI. We will not compromise sensitive information simply for the sake of speeding up a task.

For Boards, it is worth asking:

Which AI tools are approved by the organisation?

Whether directors and employees are actually using those tools rather than their own personal accounts?

What happens to information once it is entered?

Who is responsible for making sure those boundaries are understood?

Having an enterprise AI platform available does not automatically mean everybody is using it. That is where governance becomes practical rather than theoretical. It is not enough to know that AI is being used, Boards need to know where information is going when it is.

You checked the minutes. What about the Board papers?

When organisations think about AI and meeting records, the minutes are often the first place they look. However, AI may have touched the process much earlier.

A Board paper may have been drafted with AI assistance, or a report summarised before it reached the agenda. A recommendation may have been reworded or condensed before directors ever saw the original material. That does not automatically make the document unreliable, but it does mean the same level of scrutiny should apply to the papers going into the meeting as the record coming out of it.

Sometimes the signs are subtle: Americanised spellings appearing unexpectedly, the tone changing between sections, a phrase sounding unusually generic or unlike the person who normally writes the paper.

The more useful question is whether the final document has still been properly reviewed. That means checking it reflects the author’s intended meaning, confirming the facts are correct, and being alert to anything that has been added, softened or overstated along the way. It also means the person signing off the paper has actually checked the version that reached the Board, not an earlier draft. AI review should not stop at the minutes. The Board needs confidence in the information it is being asked to consider.

Has AI actually saved you time?

One of the biggest attractions of AI is the promise of efficiency. However in our experience, an AI-generated set of minutes should not be treated as a reliable first draft without significant human input and judgement.

Even getting to a usable transcript can be challenging. The recording needs to be clear enough for the technology to distinguish speakers and follow the discussion, which is not always straightforward in a Boardroom, where people may sit at different distances from the microphone, speak over one another or join remotely. Some meetings also cannot, or should not, be recorded at all.

And even when the transcript is strong, we have still seen AI:

mix up attendees and attribute comments to the wrong person

struggle to recognise when somebody joins or leaves the meeting

confuse separate agenda items or move information into the wrong section

turn something discussed into something that appears to have been agreed

lose context and nuance

introduce details that were never actually said

These are not always obvious errors. The output can read well and still be inaccurate. That is why human review needs to involve more than proofreading. Someone has to understand the meeting well enough to recognise when the meaning has changed.

When an internal employee takes on minute-taking with AI in-house, the time it takes can grow well beyond what was expected, prompting the tool, checking the transcript, cross-referencing the recording and correcting the output until it is usable.

An independent minute taker may have a visible cost whilst an internal employees time spent doing the same job can much more easily blend into the background, even when it adds up to more.

AI can absolutely create efficiencies, but only when it is used in the right way. The technology may save time. The skill of minute-taking is still what makes the record reliable.

Final thoughts

For us, AI is most useful when it supports an experienced person rather than attempts to replace the skill itself.

A professional minute taker still needs to understand what mattered in the room, where discussion became decision, which actions genuinely arose and what context needs to be preserved. AI can help with parts of that process, supporting drafting, editing, searching and summarising when used carefully. However, the judgement still needs to sit with the person creating the record.

That is where the real efficiency comes from, not from asking AI to do the job and then spending hours correcting it, but from using technology to support expertise that is already there.

So perhaps the question is not simply “Are we using AI?” It is: ‘is AI genuinely saving us time while still protecting accuracy, security, judgement and good governance?’