Artificial Intelligence: The Emergence of Frontier Firms

Microsoft released its annual Work Trend Index Report this week and it concluded: ‘‘The question isn’t if AI will reshape work – it’s how fast we’re willing to move with it’’. And they have data that shows how AI is being used more frequently in organizations.

AI might be reshaping the future of work but that doesn’t mean Copilot is doing the reshaping. People are using ChatGPT, Perplexity, and other tools, regardless of whether they have been officially sanctioned. Copilot isn’t likely having an impact on many organizations because not enough of them are paying for licenses. At least, not yet.

The emergence of Frontier Firms

Microsoft says in its Work Trend Index Report 2025 that the data suggests the emergence of ‘Frontier Firms’:

The data and insights point to the emergence of an entirely new organization, a Frontier Firm that looks markedly different from those we know today. Structured around on-demand intelligence and powered by “hybrid” teams of humans + agents, these companies scale rapidly, operate with agility, and generate value faster.

Not unsurprisingly, Microsoft predicts that within the next 2-5 years, all businesses will be on their way to becoming a Frontier Firm.

82% of leaders say this is a pivotal year to rethink key aspects of strategy and operations, and 81% say they expect agents to be moderately or extensively integrated into their company’s AI strategy in the next 12–18 months. Adoption is accelerating: 24% of leaders say their companies have already deployed AI organization-wide, while just 12% remain in pilot mode.

Microsoft wants you to act now and to decide how you will adapt. And if you read between the lines, the message is that if you fail to adapt, you’ll be left behind as other organizations reap the benefits of AI.

An AI confidence crisis

Copilot and ChatGPT often produce answers that include hallucinations and results based on information out of bounds of the context that they were asked to work with. For instance, if I ask Copilot to look at a column in a table on a Loop page, and I want to query that column only, it doesn’t confine itself to the indicated column but will happily go off looking at the same column on different pages as well.

I asked Copilot to come up several possible YouTube video titles for the last episode of This Week in IT using the show notes in Loop. As I used the same table for every episode, Copilot looked in the column on different pages from the last few episodes and came up with several video titles for previous episodes in addition to the current episode.

I got my answer in a roundabout way but Copilot provided plenty of information in its output that wasn’t relevant to my query. You could argue that I should have specified only to use information from the current Loop page. And indeed, that is a solution for the problem I experienced above. With a simple query, it’s easy to check the result and tweak the prompt as necessary. But with large complex queries, that’s not so easy.

My task for Copilot was simple. And if I can’t trust it to look in a table column on the page to give me a result, then how can I trust the Copilot Analyst agent to determine what market segment I should be targeting a new product at unless it sticks to the data and context I’ve provided? And that’s the problem with AI now. Trust. AI makes things up and it looks in places where it wasn’t asked to go.

Microsoft is working to address this in Copilot to improve adoption. Not that trust is just a Copilot problem, it’s an AI problem that exists across the board. But without trust and confidence, who is going to ask AI to make critical business decisions?

Become an agent boss

Microsoft says that we’re going to be managing agents in the future, and I think that’s probably how this will pan out. But is AI reshaping the future of work? Potentially, but very slowly. Copilot agents aren’t trustworthy enough today to perform complex tasks that are business critical. At least not without a lot of oversight.

Microsoft is waking people up to the potential of AI with the bold claim in their latest report. They want you to buy Copilot licenses but it’s expensive. And organizations want to see more evidence that they’ll be able to get a good return on investment if they spend this money.

Start off small with AI

Today, businesses should trial Copilot with a small set of employees that could benefit from AI. Users need to know how to use it, understand what it can do, and what its limitations are to make good use of it. It’s not for everyone.

But it’s still early days for AI. Let’s have this conversation again in a couple of years and it might be a different story. I would give Copilot a two and a half out of five in a performance review. It needs to up its game before I can trust it with complex, business critical tasks.