There’s a moment happening in organizations across every industry right now. Maybe you’ve witnessed it. Maybe you’ve been in the room when it happens.
The announcement comes from leadership: the organization is adopting a new AI platform. The implementation timeline is set. Training sessions are scheduled. FAQs are drafted. Communication plans are activated. Rollout begins.
And somewhere in the middle of that process, in a hallway conversation, a Teams thread, or a one-on-one with a manager, someone says something that doesn’t make it into any adoption dashboard or project status report:
"Nobody asked us."
Not a complaint, exactly. Not resistance in the conventional sense. Just a quiet, accurate observation: the people who will live with this change every day were not part of designing it.
This is the participation gap. And it may be the most underexamined reason AI transformations struggle to take hold.
The Change Built in a Room
Most AI transformation strategies are architected at the top. That’s not an indictment; it makes structural sense. AI strategy involves procurement decisions, data governance, security review, and competitive positioning. These are conversations that genuinely belong at the leadership level.
But somewhere between the executive decision and the front-line rollout, something critical gets dropped: the perspective of the people who know the work most deeply.
The accounts payable team that knows exactly where the exceptions live. The customer service reps who know which calls can never be templated. The analysts who have built workarounds over years that would break under the new system. The nurses, the underwriters, the field technicians: the people whose daily experience will ultimately determine whether the AI adoption succeeds or fails.
These are the people who are almost never in the room when the strategy is built.
And when the rollout arrives, they don’t experience it as transformation. They experience it as something being done to them.
Why We Default to Top-Down
It’s worth pausing here to emphasize that the default to top-down planning isn’t malicious; it’s structural.
AI moves fast, and competitive pressure is real. Boards and executives are watching adoption metrics. There’s a reasonable fear that broad participation will slow things down, introduce too many conflicting opinions, or derail momentum entirely.
There’s also a subtle assumption embedded in many AI rollouts: that the technology’s value is self-evident, and that resistance is simply a communication problem. If people just understood what this tool could do, they’d get on board. So the solution is better messaging, clearer FAQs, and more training sessions.
But that frames the wrong problem. Resistance to AI is rarely about a lack of information. It’s about a lack of agency, and a well-founded sense that the organization has made decisions about employees’ work without them.
What's at Stake
The most obvious cost of the participation gap is adoption. When employees feel that a change was done to them rather than with them, they comply minimally and wait it out. They use the tool when required and revert to old habits when no one is watching. The adoption metrics look passable; the actual behavior change doesn’t materialize.
But there’s a less obvious cost that rarely gets measured: organizational learning.
When front-line employees aren’t involved in AI strategy, organizations lose access to the richest source of knowledge they have — the situated, practical, hard-won expertise of the people doing the work. They don’t just lose buy-in. They lose signal.
The workaround that took a team three years to develop often encodes a real insight about why the standard process doesn’t work. The edge case that a customer service rep flags might reveal something fundamental about how customers actually use the product. When these people aren’t asked, that knowledge doesn’t disappear; it just never makes it into the design.
The result is AI implementations that are technically sound and practically brittle. They work in the use cases leadership imagined and fail quietly in the ones they didn’t know to ask about.
Enter Humble Inquiry
The late Edgar Schein, the MIT Sloan organizational psychologist whose work on culture and change shaped a generation of leadership thinking, spent much of his later career making a deceptively simple argument: that leaders ask too little and tell too much.
In Humble Inquiry (first published in 2013 and updated with his son Peter in 2021), Schein distinguishes between “telling” cultures, where expertise and authority flow downward, and inquiry-based relationships, where leaders genuinely ask questions they don’t already know the answers to. Not performative questions, or questions designed to build rapport before delivering a pre-formed answer. Questions that reflect real curiosity about how someone else experiences their work.
In a change management context, this is not a soft skill. It’s a strategic discipline. It’s the difference between learning that your AI tool can’t handle an exception your team processes 200 times a month while there’s still time to fix it, and learning it the hard way after launch.
Applied to AI transformation, humble inquiry looks like a set of questions that most organizations might not think to ask before they deploy:
- What parts of your work feel most vulnerable to this change?
- Where do you see the real potential — in ways we might not have thought of?
- What would need to be true for this to actually work for you?
- What do you know about this work that we should understand before we build the implementation plan?
These aren’t focus group questions or survey items appended to a communication plan. They’re the foundation of a design process that treats the people closest to the work as co-architects.
What Participatory AI Change Looks Like
Participation doesn’t mean consensus. It doesn’t mean every employee has veto power over the AI strategy, or that rollout timelines grind to a halt while every opinion is incorporated.
What it does mean is that employees have a meaningful role in shaping how the change happens in their context, and that their input visibly influences decisions.
In practice, this can show up in many ways:
Involving Front-Line Employees in the Pilot
Before broad rollout, identify a small group of employees who are willing to experiment. It’s great to involve champions who are already enthusiastic, but it's equally as important to involve skeptics as well. Give them real problems to work on with the new tool and ask them to document what breaks, what’s missing, and what’s surprising.
- One mid-size manufacturer piloting an AI production-scheduling tool started with six line supervisors: three enthusiastic early adopters and three vocal skeptics. They ran two weeks of live orders using the AI tool. The skeptics surfaced what could have been a key point of failure: the tool ignored the changeover time between two product lines, something the champions never thought to test. Since the issue was caught in piloting, the team was able to resolve it before deploying to production.
Making Input Visible
One of the fastest ways to rebuild trust in a change process is to show people how their feedback shaped decisions, even when the answer was “we heard you, and here’s why we went a different direction anyway.” The act of being heard matters independently of whether every suggestion is adopted.
- A logistics company did this with a simple shared “You said, we did” page on its intranet. Every piece of pilot feedback was listed next to the decision it drove, including the requests that were declined and why. Employees checked it more often than the training materials.
Asking Before Designing
Before the implementation plan is finalized, conduct structured listening sessions to understand how the work gets done. What Schein would call “access inquiries,” aimed at accessing the knowledge of those who knows more than you do about a specific domain.
- At one B2B sales organization, an afternoon spent shadowing account reps revealed that the “standard” deal flow the vendor had designed around almost never happened in practice; reps were constantly looping back through procurement and legal. That single session reshaped the entire configuration and resulted in a workflow that actually made sense for how the team operated.
Building in Ongoing Feedback Loops
Participation isn’t a one-time event at the beginning of a rollout. The most durable AI implementations treat adoption as a continuous conversation, with a structure for employees to flag what’s working and what isn’t, and visible evidence that those flags are being read.
- One operations team implemented a standing 30-minute monthly review where flagged issues were triaged live and assigned owners in front of everyone. Knowing a real person would read and act on each problem, rather than disappearing into an unmonitored inbox, kept the suggestions coming.
The Question Underneath the Question
When an employee says “nobody asked us,” they’re not just expressing frustration about process. They’re surfacing a deeper question about trust: Does this organization actually value what I know about my own work?
In the context of AI transformation, where employees are often wondering whether the organization is building something that will ultimately replace them, that question carries extra weight. The participation gap lands differently when the stakes feel existential.
Leaders who close the participation gap aren’t just improving adoption metrics. They’re sending a message that’s more important than any FAQ or all-hands presentation: we believe your expertise matters to how this goes.
That message, more than any feature of the technology itself, is often what determines whether AI transformation becomes something people endure, or something they own.
Works Cited
- Schein, Edgar H., and Peter A. Schein. Humble Inquiry. 2nd ed. Berrett-Koehler Publishers, 2021.