LEGO® SERIOUS PLAY® for AI Adoption: How to Get Teams Talking Honestly About AI at Work
AI adoption is the process of putting artificial intelligence tools into everyday work and reorganizing roles, decisions, and outputs around them. Most organizations are further into that process than their people are willing to say out loud. Mercer's Global Talent Trends 2026 survey of 12,000 workers and business leaders found that 40% of employees now fear losing their job to AI, up from 28% two years earlier. Microsoft and LinkedIn's Work Trend Index found that 52% of people who use AI at work are reluctant to admit using it for their most important tasks.
LEGO® SERIOUS PLAY® is a facilitated workshop method in which every participant builds a three-dimensional model in response to a question, then explains it to the group. Applied to AI adoption, it moves the conversation off the defensive ground of "is my job safe" and onto a shared table where people model what their work actually consists of, which parts of it are changing, and what they want to keep human. This article covers why AI rollouts stall on fear rather than technology, how a half-day LEGO® SERIOUS PLAY® AI adoption workshop is structured, how to choose a facilitator, and how to measure the result at 30, 60, and 90 days.
Direct answer: LEGO® SERIOUS PLAY® helps AI adoption by making people describe their actual work before anyone discusses automating any of it. Instead of a leadership team presenting an AI strategy on slides, every participant builds a model of what their role really involves, and the group then builds one shared model of how the team will work with AI inside it. The output is a set of agreed working principles — what AI does, what stays human, and what people will say openly — that the people who have to live with them helped construct.
Why AI Rollouts Stall on Fear, Not Technology
The tooling is the easy part. Licenses get bought, pilots get run, and a dashboard somewhere shows seat activation climbing. What cannot be bought is the willingness to say what the tool is actually being used for.
What does the data on employee AI sentiment actually say?
The numbers describe a workforce that is already using AI and quietly hiding it. Anxiety and adoption are rising together, and training is falling behind both.
- 40% of employees fear losing their job to AI, up from 28% two years earlier, according to Mercer's Global Talent Trends 2026 survey of 12,000 workers and business leaders worldwide.
- 52% of people who use AI at work are reluctant to admit using it for their most important tasks, and 53% worry that doing so makes them look replaceable — Microsoft and LinkedIn's Work Trend Index, based on 31,000 respondents across 31 countries.
- 78% of AI users bring their own tools to work, per the same study, which means a large share of real adoption is happening outside whatever policy the organization has written.
- Only 36% of workers say their employer provides the training and resources they need to use AI, down from 45% a year earlier, in a Jobs for the Future survey released in March 2026.
- 18% of US employees think it is likely their job will be eliminated within five years because of AI or automation, rising to 23% among employees at organizations that have already adopted AI (Gallup, 2026).
Why does the standard AI rollout session make it worse?
A typical rollout session opens with a slide reassuring everyone that AI will augment rather than replace them. It is usually delivered by the person who would sign the headcount decision if it went the other way.
Everyone nods. Then they return to their desks, keep using the tools they were already using, and say nothing about it. The organization gets adoption without learning: no shared view of what is working, no visibility on data risk, and no honest input into the redesign it is about to attempt.
In an AI rollout, silence is not adoption. It is unmeasured usage plus unspoken fear, and neither one shows up in your license dashboard.
What Is LEGO® SERIOUS PLAY® and Why Does It Work for AI Conversations?
LEGO® SERIOUS PLAY® is a facilitated meeting and problem-solving method in which participants build models with LEGO® bricks to represent their thinking, then tell the story of what they built. It originated inside the LEGO® Group in the late 1990s, drew on constructionist learning research, and was released under an open-source model in 2010.
How does the method work?
- Pose the challenge. The facilitator asks a specific question, such as "What does a good day of your work actually look like?"
- Build. Every participant builds their own model. Nobody watches; everyone works.
- Share. Each person tells the story of their model. The group asks questions about the model, not about the person.
- Reflect. The facilitator draws out patterns, contradictions, and shared themes across the builds.
Why does building beat discussing when the subject is AI?
AI conversations fail in a specific way: people answer the question they think is safe rather than the one that was asked. The mechanics of the method change what is safe to say.
- The model absorbs the exposure. Admitting "I use a chatbot for the first draft of every client email" is a career calculation. Pointing at a brick on a model and explaining what it stands for is a description.
- Everyone contributes. Because each participant has to build and then explain, the session does not become a conversation between the two people most comfortable with the technology.
- Abstractions become specific. "AI will handle the low-value work" is unfalsifiable. A model showing which six of a person's eleven weekly tasks are in question is something you can argue with.
- The task separates from the role. Building a job as a set of components lets a group discuss automating one component without anyone hearing that their identity is being deleted.
Four AI Adoption Problems a LEGO® SERIOUS PLAY® Workshop Can Solve
The method is not a substitute for an AI strategy, a tool selection process, or a data policy. It is a way of getting accurate information into all three while there is still time to act on it.
Surfacing shadow AI use before it becomes a governance problem
Most organizations do not know which tools their people are pasting company information into. A build-and-share format gets that on the table faster than a survey, because the unit of disclosure is a model of a workflow rather than a confession attached to a name.
Separating the task from the role
People hear "we are automating parts of this function" as "we are removing you." Modeling a role as a set of discrete components — the analysis, the judgment, the relationship, the formatting — lets a team hold a real conversation about which parts change and which do not.
Deciding what stays human, on purpose
Every team has work it believes should not be automated: a particular client conversation, a specific quality check, the moment a junior person learns something by doing it slowly. If that list is never made explicit, it gets eroded by default rather than by decision.
Keeping the people who will actually make the rollout work
The employees most likely to leave during an AI transition are the ones with options and no voice in the redesign. A session in which a specialist explains their model of the work to the leaders planning the rollout is a concrete signal that their read on the job still counts.
A workforce cannot redesign work it has never accurately described. Model the job first; decide what AI does second.
How to Run an AI Adoption Workshop Using LEGO® SERIOUS PLAY®
The structure below is a half-day format of roughly 3.5 to 4 hours, for 8 to 20 participants drawn from a single function or from a cross-functional team that shares a workflow. It is designed to run either before a tool rollout or within its first 90 days, while working habits are still forming.
Click the '+' button below to explore the detailed workshop stages.
1. Skills Build & 2. What My Work Actually Is
1. Skills build and readiness (20–30 minutes): Participants complete two or three short warm-up builds to learn the mechanics: build fast, build metaphorically, explain what you made. This stage removes self-consciousness and establishes that nobody is being judged on construction quality. Skipping it is the most common reason first-time sessions underperform, and it matters more than usual here because the topic is already threatening.
2. Individual builds — "what my work actually is" (25–35 minutes): Each person builds a model of their real week: what they produce, what they decide, who depends on them, and where the time goes. Everyone then explains their model. The facilitator captures recurring components without yet mentioning AI at all. Holding the technology out of the room for this stage is deliberate — it produces a description of the work rather than a defense of it.
3. What Is Already Changing & 4. What Stays Human
3. Marking what is already changing (30–40 minutes): Participants physically mark the parts of their own model that AI already touches, could plausibly touch, or has changed in the last year. Because the marking happens on an object each person built themselves, shadow usage tends to surface here as a matter of fact rather than as an admission. The facilitator collects the pattern across the room, not the individual cases.
4. Naming what stays human (25–35 minutes): The group identifies the components that should remain human and says why — client trust, regulatory judgment, quality control, or the learning that only happens by doing the work. These are placed on the models physically. The goal is an explicit, agreed list, not a consensus that everything is precious.
5. Shared Working Model & Commitments
5. Shared model, working principles, and individual commitments (45–60 minutes): Working as one group, participants build a single model of how the team will operate with AI in it, including the handoffs between human and automated steps. Everyone must have an element of their individual build represented somewhere in the shared model. The group then converts it into five or six working principles — each phrased as something a person could actually follow on a Tuesday — covering what AI does, what stays human, and what people will disclose openly. Each participant states one specific commitment for the next 30 days. The models are photographed and the principles go to whoever owns the AI rollout as workshop output, not facilitator interpretation.
Standard AI Rollout Session vs. LEGO® SERIOUS PLAY® AI Workshop
Both approaches take a similar amount of calendar time. They produce very different information.
| Dimension | Standard rollout session | LEGO® SERIOUS PLAY® workshop |
|---|---|---|
| Who speaks | Leadership, plus the two most confident adopters | Every participant, structurally |
| Source of the work description | Process maps and job descriptions | The people doing the work, built |
| How shadow AI use surfaces | Rarely, or after an incident | In the room, marked on a model |
| Primary output | A policy document and a training link | Working principles, a "stays human" list, individual commitments |
| Main risk | Compliance without candor | Weak facilitation producing an enjoyable session with no output |
How to Choose a Facilitator for AI Adoption Work
The risk in the last row of that table is the one worth dwelling on. A LEGO® SERIOUS PLAY® session run by an inexperienced facilitator is a pleasant morning that changes nothing, and an AI conversation is an unforgiving place to discover that.
What should you ask before you book?
These questions separate practitioners from people who attended a course once. Ask them of any provider, including this one.
- How many years have you been delivering this method, and how many facilitators or sessions is that? Longevity indicates a process refined against real rooms rather than theory.
- Have you facilitated groups where people have a direct incentive not to be honest? An AI session contains one by definition. Ask how they handle an executive who summarizes on everyone's behalf.
- How do you adapt when the group is 8 people versus 40? Group size changes the entire build-and-share arithmetic. A facilitator who has only run one size will run the wrong session.
- What happens to the output after the session? If the answer stops at "photographs," the workshop will not connect to the AI rollout plan.
- What is your relationship to the method's origins? The method has been open source since 2010, so anyone may facilitate it. That makes depth of training and provenance a differentiator rather than a formality.
Why does experience level matter more here than usual?
Serious Play Business has spent 19 years focused on this single method, training more than 1,800 facilitators across 38 countries. Founder Dr. Denise Meyerson was selected as one of the original four Master Trainers of the LEGO® SERIOUS PLAY® method and has been a Master Trainer since 2010, working with organizations including Capgemini, Kaiser Permanente, Syngenta USA, the US Air Force, and Virginia Commonwealth University.
That background matters in an AI session for one practical reason. When someone marks half their model as automatable and the room goes quiet, the facilitator has to decide in real time whether to push into it or park it. Getting that judgment wrong either wastes the session or detonates it.
How to Measure Whether the Workshop Worked
Culture work gets cut when it cannot be evidenced. Attach measures to the session before it runs, not after.
What should you capture on the day?
- The list of work components the group marked as already touched by AI, in their own words.
- The agreed "stays human" list, with the reason attached to each item.
- Five or six working principles, no more.
- One written 30-day commitment per participant, with a name against it.
- Photographs of the individual models and the shared model.
What should you track at 30, 60, and 90 days?
- Day 30: Completion rate of individual commitments. This is the earliest honest signal of whether the session carried authority or was theater.
- Day 60: Disclosed AI use against sanctioned tools. If people are naming what they use without being asked, the disclosure norm took. If shadow usage is unchanged, the principles did not reach the desk.
- Day 90: Whether the "stays human" list has been referenced in an actual process or tooling decision — and voluntary regretted attrition among participants compared against equivalent teams that did not attend.
Run the workshop before the rollout, or inside its first 90 days. That is when working habits are still forming, and when a shared model can still change what forms.
Transform Your Strategy Conversations
If you are rolling out AI and need the real picture of how your teams work before you redesign it, a facilitated LEGO® SERIOUS PLAY® session will get you there faster than another round of adoption surveys. Tell us the shape of your rollout and we will design the session around it.
To understand the full methodology behind the workshop design described above, read the LEGO® SERIOUS PLAY® method explained.
Frequently Asked Questions
Should we run this before or after we roll out the AI tools?
Before is stronger, because you get an accurate description of the work while people still have no reason to be defensive about it. If the rollout has already started, run it inside the first 90 days — working habits are still forming at that point, and shadow usage has not yet hardened into something people feel they have to protect.
Will people really admit to unsanctioned AI use in a facilitated session?
More readily than in a survey, though not because the method is a truth serum. The disclosure happens as a description of a workflow on a model the person built, in a room where everyone is marking their own model at the same time. It also helps to state at the start that the session is not an audit and that the output going upward is the pattern across the room, not a list of names.
Does this replace an AI policy or a training program?
No. It feeds them. A workshop produces the ground truth — what the work is, what people are already doing, and what the team believes should stay human — that a policy and a training curriculum should be built on. Writing the policy first and consulting afterward is the sequence that produces compliance without candor.
Can this run for remote or distributed teams?
Yes. Kits are shipped to participants in advance and the build-and-share structure runs over video, with adjustments to timing and to how the shared model is assembled. Distributed teams are often the ones with the widest variation in AI practice, since there is less incidental visibility into how colleagues actually work.
Do we hire a facilitator, or certify someone internally?
Both are viable, and the choice usually turns on how often you expect to need the method. For a one-off rollout, an external facilitator brings neutrality that an internal employee — who has their own stake in the redesign — cannot. If AI transition, change, and team alignment work is ongoing, certifying an internal facilitator is generally more economical. Serious Play Business offers a Foundational Certification at $750 USD and an Advanced Certification at $1,300 USD, both fully online and self-paced.
What is a Master Trainer of the LEGO® SERIOUS PLAY® method?
Master Trainers were the practitioners originally selected and trained to teach the method to other facilitators before it was released as open source in 2010. Only four were selected originally, one of whom was Dr. Denise Meyerson, founder of Serious Play Business. Since the method is now open source, anyone may facilitate it — which is precisely why depth of training and years of practice are worth checking when you select a provider.