AI is changing work, roles and expectations faster than organisations can fully define what comes next. You cannot manufacture certainty. You can create the conditions for people to trust how the change is being led and adapt as the picture develops.
Make sense. Understand what is happening. Adapt from the evidence.
Request the one-pager See the way inYou can't automate trust.
You have to create the conditions for it, and keep earning it as the change unfolds.
AI adoption is widespread. Realised impact is not.
69% of firms report actively using AI.
More than 80% report no impact on either employment or productivity over the previous three years.
Yotzov et al., NBER Working Paper 34836, 2026. Survey of almost 6,000 CEOs, CFOs and senior executives across the US, UK, Germany and Australia.
The harder part is how the organisation learns to work differently around it. When the route is still emerging, leaders have to set direction while being clear about what is known, what is assumed and what is still unfolding.
When the destination is still taking shape, people judge the change by how it is being led.
The question people care about most is often the hardest one for leaders to answer.
Strategy can be explained from the front of the room. “What does this mean for my role and my expertise?” cannot. Research on change uncertainty finds that uncertainty about one's own job is the most stressful kind, and that it falls through participation, not announcement.
Bordia, Hobman, Jones, Gallois & Callan (2004), Journal of Business and PsychologyWith AI, that question cuts deeper. It reaches people's sense of competence, autonomy and professional identity, not only their workload.
Hermann, Puntoni & Morewedge (2026), Harvard Business ReviewWhat leaders usually see
Plans, milestones, programme status, usage data, training completion and formal feedback.
What can remain hidden
Different interpretations, quiet doubt, threatened expertise, loss of agency, defensive compliance and questions people do not feel safe or able to ask.
In these conditions, leadership is less about projecting decisiveness and more about adaptive navigation: enough direction to move, enough honesty to admit uncertainty, and enough trust for people to learn and adjust alongside you.
Separate what is known, assumed, undecided and still emerging. Being clear about uncertainty is more credible than confidence without evidence.
Surface where leaders genuinely agree and where they see the future differently. Then involve people in what the change means for their work, rather than relying on the cascade to carry it.
Trust is built in how leaders handle uncertainty, challenge, bad news and decisions that cost them something. A 2026 study found that leaders who gave people room to raise concerns about AI, and were open to being influenced, strengthened trust in themselves and, through it, trust in the technology.
McGuire, De Cremer & Narayanan (2026), Journal of Business ResearchWhen the evidence shows the conditions are getting in the way, change the conditions. Transformations run as fixed-goal projects tend to double down when early results disappoint; those run as learning journeys adjust.
Kaganer & Loch (2026), Harvard Business ReviewWhere the real constraint is structural, leadership behaviour alone will not resolve it. But leaders are still the people who have to see it clearly and decide what changes.
You cannot lead what you do not understand. Leaders experience transformation from one position. The people being asked to adapt experience it from another. Start with the leadership team, then understand how the same change is being interpreted and experienced across the organisation. Diagnosis comes before intervention, so effort follows the evidence rather than habit.
Clarify the direction of travel, surface the assumptions leaders are making about the people affected, and separate what is known from what is still uncertain.
Meet people where they are. Use structured dialogue and evidence to understand how the change is actually being interpreted and experienced.
Act on the conditions that are genuinely in the way. Test the response, return to the evidence, and adjust.
Start with leaders. Follow the evidence. Change the conditions.
HuSAT, the Human System for Adaptive Trust, is our framework for reading the conditions that shape how people adapt to AI and significant change. It shows what is happening beneath the adoption numbers, without reducing people to a single readiness score.
Sensemaking. Trust and safety. Agency. Fairness. Professional identity.
Adaptive trust is not certainty about where the transformation will end. It is confidence that the way it is being navigated is credible enough for people to keep engaging, questioning and adapting, and for trust to be tested and rebuilt as the change and the technology evolve.
A focused, discussion-led working session for leadership teams sponsoring AI, automation or other significant change. It makes the current position explicit, surfaces assumptions and alignment gaps, and builds a clear view of the leadership environment before any workforce intervention.
Leaders leave with one condition worth protecting, one assumption worth testing against evidence, and one behaviour or decision worth changing. And a clear recommendation on the next step, which may be workforce diagnosis, deeper leadership work, a targeted intervention, or nothing further from us.
Want the one-page overview? Request the Centaur Dynamics one-pager.
We do not assume leaders or employees should already have a settled view of a future that is still emerging.
We test leadership assumptions against what the workforce actually reports, and let the diagnosis choose the intervention.
We treat caution, silence and dissent as information about the system, not behaviour to correct.
Independence creates room for candour that formal reporting can miss. People can describe what they are actually experiencing, not what the programme expects to hear.
The approach combines behavioural science with experience inside complex, regulated, AI-enabled transformation.
If your leadership team is leading AI or significant change while the destination is still taking shape, start with the one-page overview of the approach. It is available to download as soon as you submit.
Tell us briefly where things stand. It helps make any follow-up conversation useful from the first minute.
No pitch, and no diagnosis sold before we know you need one.
James will follow up personally, not a sales team. The first conversation is simply to understand where things stand and whether there is anything useful here.
James, Centaur Dynamics
Prefer to talk first? info@centaurdynamics.com