Microsoft published its 2026 Work Trend Index last week. I’ve compared it against the 2025 one, and the thing that struck me wasn't either report on its own. It was the shape of the conversation between them.
In 2025, the headline was "buy intelligence on tap". Agents as digital labour. Frontier Firms scaling capacity. The whole thing read as a sales pitch dressed up as research, which is fine because that's basically what these reports are. But underneath the marketing layer there was a genuinely useful observation: 82% of leaders said this was a pivotal year to rethink strategy and operations, and 80% of the workforce said they didn't have the time or energy to do their existing work.
Roll forward twelve months. The 2026 report has shifted the frame entirely. The opening line is "as AI and agents take on execution, our own agency expands". Karim Lakhani's foreword talks about the operating model of the firm, not the tool. The central data point is that organisational conditions, things like the culture leaders set and how managers actually support their people, account for roughly twice the AI impact of individual effort and mindset combined.
Which is essentially Microsoft saying out loud what most of us working with senior teams have been saying since the tools came out.
The technology was never going to be the hard part. The hard part was always going to be everything that has to change around the technology for it to actually compound.
What changed between the two reports
A few things worth pulling out, because they tell you where the conversation is going.
The 2025 report was structured around what individuals and firms would do with agents.
The 2026 report is structured around the gap between what employees can do and what their organisations are built to let them do. Microsoft calls this the Transformation Paradox:
- 65% of AI users fear falling behind if they don't adapt with AI quickly,
- 45% say it feels safer to focus on current goals than to redesign work with AI,
- only 13% say they're rewarded for reinventing work even when it doesn't immediately produce results.
That's the trap. The very same operating systems that made the business successful in the first place are now the things actively holding it back from absorbing the technology that could grow it further.
Microsoft also introduced the idea of a Learning System: a firm that captures the signals from agent workflows and feeds what worked, and what didn't, back into how the business operates. They're calling the output of this "Owned Intelligence". Institutional know-how that compounds. I think this is the most useful idea in the 2026 report and it deserves more attention than it's getting. Most organisations are still treating AI adoption as a procurement exercise. The ones pulling ahead are treating it as an internal R&D programme.
Here's where I find myself nodding and disagreeing at the same time.
Nodding because: yes, this was always going to be the shape of it. Microsoft’s own AI Event that I attended in February had the same core message running through it and it’s same one I've been giving clients for months. The focus has to be on the people using the tools and the people coordinating between them, and on whether leaders and managers create the conditions for that to actually happen inside the business. Without that, the tools just become another shelf of underused software.
Disagreeing because: I think the report still slightly under-states how hard this is for established organisations. The bigger the company, the more predefined the processes, the more entrenched the culture, the more layered the hierarchy. None of those things were designed to adapt to a technology that changes every fortnight. Is it really surprising that adoption is uneven? An AI-native startup with 12 people doesn't have to renegotiate eight years of operational habits to ship a new agent. A 600-person insurance broker does.
Which is why I think the AI Maths Framework I've been using with clients keeps holding up under the data.
The framework starts from a simple observation. There are four operations a company can be performing on its people with AI right now, and three of them are start points, not destinations. Division is when AI lands as fear, cost pressure, and extraction. People are watching colleagues leave and wondering if they're next. Subtraction is when AI has stripped low-value tasks out of the day but the work itself hasn't really changed. Roles feel a bit thinner. Addition is when the tools are present and being used, but they're sitting on top of an operating model that was designed for a different era. Most SMEs I work with are somewhere between Subtraction and Addition. And then there's Multiplication, which is the only real destination. It's where AI is helping people and teams do work they genuinely couldn't do before. Not the same work faster, but different work.
The important thing about reading the framework that way is that Division, Subtraction and Addition aren't a ladder you climb in order. They're snapshots of where a business currently sits with AI, and each one needs a different first move. A firm stuck in Division has to deal with the fear and the extraction story before anything else gets a fair hearing. A firm in Subtraction needs to start asking what the freed-up time is actually for. A firm in Addition has to redesign workflows, not procure more tools. The destination is the same in all three cases. The route in is what changes.
Now read the Microsoft 2026 data through that lens and the shape of the problem comes into focus. The 50% sitting in the "Emergent" zone, with individual practice and organisational conditions both still taking shape, are mostly start points of Addition. Tools have arrived, nothing structural has moved underneath them. The 10% in "Blocked Agency" are capable employees in firms that are stuck in Addition and refusing to redesign. The 16% in "Stalled" are sitting in Subtraction or Division and treating that as the strategy rather than the starting line. And the 19% in the Frontier zone are the ones actually at Multiplication. They're not there because they bought better tools. They're there because their managers openly use AI (85% vs 64%), set quality standards (83% vs 57%), create space for experimentation (84% vs 61%), and reward reinvention even when results aren't met (26% vs 11%).
That last one is the tail to pin on the donkey:
Reward reinvention even when it doesn't work.
That's a permission and psychological safety structure, not a technology choice.
The route from any of those start points to Multiplication is what I've been calling the Multiply Method with clients. It's not a programme of tool rollouts. It's a sequence that begins with diagnosis (where are you really, not where do you think you are) and ends with making the new way of working durable enough that it doesn't fall apart the moment the original champions move on. I won't unpack the whole thing here but the headline is this: the gap between Addition and Multiplication isn't bridged by buying more, it's bridged by redesigning work.
What I'd take from this if I were running a senior team right now
A few things I'd be paying attention to.
- First, the data point that should be keeping leaders up at night is buried a bit deeper than the productivity numbers. 65% of your AI-using employees are anxious about falling behind, and 45% of them are choosing the safer path of doing their existing job better instead of redesigning it. Which is a culture problem dressed up as a skills problem. Skills you can train. The fear of being punished for trying something new and having it not land is a much harder thing to shift.
- Second, the Owned Intelligence idea is worth taking seriously. Every time an agent runs a workflow in your business, it generates a signal: what worked, what failed, where the human had to intervene, what the customer said about the output. If those signals stay trapped in individual workflows, you've automated execution and learned nothing. If you capture them and feed them back into how the business operates, you've built something that compounds. That's the actual moat being built, and it's not the same as having "an AI strategy".
- Third, the role of the manager has quietly become the most important variable in the whole picture. The 2026 data shows a 17-point lift in reported AI value when managers actively model AI use themselves. Not when they mandate it or when they pay for tools. But when they actually use it visibly in front of their teams. I think a lot of senior leaders have outsourced the AI question to a Chief AI Officer or an IT lead, and the data suggests that's the wrong move. The judgement about how AI gets used inside a team belongs to the person who runs that team.
- Fourth, and this is the one I'd put most weight on: stop treating AI as a cost-reduction story. Some of it will reduce cost, and that's fine. But if the only question being asked at board level is "where can we use this to cut headcount", you're locked into Division. The Frontier zone firms aren't there because they got rid of more people faster. They're there because they figured out how to redesign work so that the human judgement and expertise that made the business successful in the first place now gets applied to bigger and harder problems. That's a different question entirely.
Where my head’s at
Two reports, twelve months apart, telling the same story from different angles. The 2025 one said: the tools are here, go and scale them. The 2026 one stepped back and said the tools were never the problem in the first place, the organisation is. I think both are true, and I think the 2026 framing is the more useful one because it puts the responsibility back where it actually belongs.
Where your firm starts from, whether that's Division, Subtraction or Addition, doesn't really come down to how good your AI is. And whether you get to Multiplication doesn't come down to it either. It comes down to how your leaders show up in front of their teams, how your managers create conditions for people to try things, how your culture treats reinvention when it doesn't immediately work, and whether you're willing to redesign work around what's now possible rather than retrofitting AI to what already exists.
I don't think this is easy. I do think it's the actual job.
If you're reading this and finding yourself in Subtraction or Addition, that's where most SMEs are. The interesting question isn't whether you're "ahead" or "behind". It's what an honest first move toward Multiplication would look like from where you actually are.
Enjoy this? ♻️ Repost it to your network and follow Martin Wheatley for more.
✅ I help senior leaders close the gap between knowing AI matters and knowing what to do about it, starting with decision-making, not technology.
🧭 Get your copy of the Multiply Method at www.theaileaderlab.com/roadmap
It's a FREE roadmap walking you through how to multiply the capabilities of your people and your business with AI.
🗣️ If that's the conversation you want to have, book a call here: https://calendly.com/martin-rawsonellis/30min
📍 Or visit: www.theaileaderlab.com
Sources:
Microsoft 2026 Work Trend Index: https://news.microsoft.com/annual-work-trend-index-2026/#
Microsoft 2025 Work Trend Index: https://www.microsoft.com/en-us/worklab/work-trend-index/2025-the-year-the-frontier-firm-is-born