For years, most boards have followed the same ritual. Two weeks before a board meeting, directors receive a thick digital packet (often over 100 pages) and do their best to find the “needle in the haystack” before showing up in the boardroom. Florin Rotar, group CTO at Atos and former chief AI officer at Avanade, believes that those days are rapidly coming to an end.
“I think it's important to move away from the old-school approach of big decks and piles of materials in terms of being prepared before the board meeting, asking the right questions during the board meeting, and making sure there's good follow-up,” Lothar says. “It really helps you digest what’s important.”
A veteran technologist and board advisor, Rotor sees AI not as a replacement for boards, but as an amplifier. AI will soon become important. “I don't think anything should be automated. I think it should be scaled,” he says. “With AI, boards can have a richer perspective, see the needle in the haystack, and support diverse thinking. This is very powerful, and I think this will become the norm.”
From board book to “virtual board member”
The cutting-edge boards Rotar is working on already use AI to summarize sprawling pre-reads, surface KPIs and financial anomalies, and flag potentially hidden trends. “By being able to digest information more quickly and integrate information, we can find out through the data what really matters, which is the key here,” he says.
Some are trying to push it even further by using AI as a simulation partner. “This is experimental for now, but I've seen it used as a sparring partner in simulation scenarios,” says Lothar. “It's a bit like a tabletop exercise of how cyber used to be done. We use AI to simulate and prepare for scenarios like activist investors and unilateral takeover attempts, and we use AI to provide a contrarian perspective.”
He said it's not hard to imagine AI playing a more visible role in deliberations. “AI will be used to augment boards, to provide additional perspective, to help with a bit of role-playing and what-if analysis,” he says. “Who knows? We'll probably have AI agents as virtual board members at some point.”
He noted that some sovereign wealth funds in the Middle East are already publicly experimenting with “virtual board meetings” to support investment decisions. “I’m sure it’s a bit of marketing and hype,” he says with a laugh. “But I don't think it's completely outside the realm of possibility.”
Sovereignty AI as the “corporate cortex” of the board of directors
Behind the scenes, technology itself is also changing. Lothar said serious efforts are moving away from a one-size-fits-all publishing model to tightly controlled and highly customized systems.
“For serious stuff, we usually use customized versions, because otherwise there's too much risk,” he says. “There will be hallucinations, and you'll have to worry a little bit about security and privacy.”
On the frontier, some companies are going even further. “In the most sophisticated scenario I've ever seen, it's not just customization, it's a custom, proprietary model, it's a sovereign model,” Rotor explains. “Today, what companies build almost from scratch is part of their secret sauce, their most important asset. It's know-how, the 'corporate cortex,' if you can use that word.”
Just a few years ago, this would have been out of reach for all but the largest companies, he notes. “What was $1 billion five years ago and $100 million two or three years ago is suddenly about $10 million,” he says. “It's rapidly becoming more affordable and more realistic. I'm not saying it's easy for the average midsize business, but it's doable.”
A new kind of board work
If boards want to see real value from AI, they need to change the way they personally handle AI, Lauter says. It starts with using systems that are actually capable of deep inference and pushing them to the hilt.
“There is no substitute for practical experience,” he insists. “You can join every school program in the world. Nothing beats the experience of using AI on a daily basis.” Pitfall: “Unfortunately, most people are using free or degraded versions of AI. My recommendation is simply to invest in purchasing access to the best AI that money can buy.”
He also urges board members to treat AI less like a magical oracle and more like a very unconventional colleague. “To get really good results with AI, especially modern inference AI, you have to approach it in much the same way you would approach a human, but some aspects of it are very atypical of what a sane human would do,” he says. “The more active you are with AI, and the more you tease it, speculate with it, and question it to some extent, the better it will become.”
Rotar also allows you to use multiple models for the same question and critique each other. “The best results I get are when I ask them to do pretty much the same thing and then pit them against each other and ask them to critique and synthesize each other's work,” he says. “You would never do that to a human. It would show a lack of trust. But that's actually how AI works.”
