Nearly eight in ten employees worldwide say they trust their employer to do what is right, 78 percent according to the 2026 Edelman Trust Barometer. That sounds reassuring, until you look at who feels that trust tested hardest during an AI transformation. Not frontline staff, not the boardroom, but the layer in between. Research from McKinsey shows that middle managers are the most likely to feel anxious about AI-related change at work: 25 percent, compared with 20 to 21 percent across every other level, from frontline to the executive team. Precisely the layer that stands between strategy and daily execution.
What happened
McKinsey's People & Organizational Performance Practice published "AI transformations run on trust" in August 2026, based on the AI Individual & Organizational Readiness Assessment Panel Survey, conducted between February and April 2026 among 750 English-speaking respondents. The central finding: trust in the organisation and its leaders is one of the strongest predictors of AI readiness. Organisations with high trust levels extract more value from their AI initiatives than organisations where that trust is missing.
The anxiety figures make this concrete. Employees with low trust in their organisation's support during an AI transformation are one and a half times more likely to feel anxious than employees with high trust: 29 percent versus 19 percent. And the middle manager stands out at 25 percent, higher than the frontline, higher than senior management, higher than the executive layer.
McKinsey also warns against a tempting short-term reflex: hollow reassurance. Lines such as "this is not about cost cutting, no jobs will be lost" feel good in the moment but build no trust. Worse, they break trust the moment a later action contradicts that promise, for instance the first time AI does eliminate a role.
Why this matters
Previous waves of technology could be rolled out top-down. You bought the software, trained people on the buttons, set a deadline. AI transformation does not work that way. Value only emerges when employees experiment themselves, recognise use cases in their own work, and dare to redesign workflows. That requires people who feel safe enough to take it. Anxiety is precisely what holds that back.
This is where it gets difficult at the middle manager level. That layer translates strategy into daily execution, leads teams who must themselves experiment, and must simultaneously model the behaviour it asks of others. A middle manager who is uncertain about their own role in the transformation inevitably carries that uncertainty into every team meeting. An organisation that neglects this layer in a transformation effort neglects the hinge point the entire movement depends on.
McKinsey identifies four actions that build trust, none of which centre on a grander communication plan.
Offer a plan, even an imperfect one. Communicating honestly about what you know, what you do not yet know, and how decisions get made works better than waiting until everything is settled. That also means communicating about setbacks, and making visible what pilot projects are teaching and how those insights shape the course. Admitting uncertainty builds more trust than selling false certainty.
Go into the organisation yourself. Senior leaders often get an incomplete picture, because signals get filtered, softened, or simply lost on their way up. Team visits, sitting in on meetings, walking alongside the people actually doing the work: that yields sharper information than a dashboard. Especially with middle managers, since they see both the strategic intent and the daily reality most clearly. Those who only listen upward never hear what gets lost along the way.
Invest in people, continuously rather than once. A single AI course is not enough. What works is learning in the flow of work, coaching, and role-specific support: someone returning from parental leave needs different help than a deep technical expert. This also applies when roles are lost: fair severance, coaching, time. Some companies keep employees on payroll and benefits between assignments rather than severing the relationship immediately. Capability building is not a project with an end date, it is an ongoing responsibility.
Equip a broad layer of leaders to carry trust. Trust does not scale if it stays with a handful of senior leaders. Employees experience a transformation through their direct manager and their colleagues, not through the CEO. This works through discussion guides, practice sessions for difficult conversations, peer coaching, and above all the recognition that not every leader already knows how to do this. Trust that stays at the top is not trust that reaches the floor.
Those four actions do not sit side by side, they reinforce each other. That is where the Rescope Transformation Platform steps in, launched recently by Rescope out of its own experience working with clients and a conviction about what an AI transformation needs in order to hold together. It is striking to see that independent research from McKinsey now describes exactly the same four actions as what makes or breaks trust. The platform does three things at once.
One: for everyone in the organisation, using competency matrices and linked training, it makes crystal clear what is expected and where someone stands today, not a loose menu of courses but a learning path built for the role. For the new roles an AI transformation brings with it, such as Champion, Transformation Lead and AI Engineer, that learning path goes a step further. That is the ongoing investment in people McKinsey calls for, and it does not stop at the boardroom: the Transformation Lead gets the same learning path as the Champion, not a diluted version of it.
Two: it brings all of the organisation's use cases together in one managed pipeline, from first idea to production, with a clear owner and a visible ROI, instead of scattered experiments springing up across the organisation that never come together. That is the plan McKinsey asks for: not finished, but visible and current, rather than a message sent once and then left to go quiet.
Three: it gives the Steerco its own view that brings together the progress of capabilities, use cases and ROI, with a separate space for the strategic bets the executive team steers itself. That way the same visibility runs from the shop floor to the boardroom, the hinge point this article is about, made literally visible all the way up to the Steerco.
The real lesson
It is tempting to treat trust as a communications problem: the right message, at the right time, from the right spokesperson. But trust is not created by one town hall or one internal email. It is created through hundreds of small decisions over time: which question a manager takes seriously, which setback they admit rather than hide, which employee they bring along rather than walk past.
The middle manager sits at the centre of this, not by coincidence. They are simultaneously the most anxious and the most decisive for whether those hundreds of decisions land well. An AI roadmap that skips this layer and communicates straight from the boardroom to the workforce misses the hinge point where the transformation either holds or breaks.
Where does that hinge point sit in your organisation today, and who is equipped to carry the trust you expect it to hold?