The First Thing We Protect Is Your People

Every AI engagement starts with a fear that nobody talks about in the kickoff meeting.

The person in finance who’s been reconciling reports for 14 years hears “AI strategy” and thinks “they’re automating my job.” The mid-level manager who championed the last technology initiative that failed hears “transformation” and thinks “I’m going to be blamed again.” The IT director who’s been holding the systems together with duct tape and institutional knowledge hears “modernization” and thinks “they’re going to find out how bad it really is.”

These are rational responses. These people have watched consultants walk in before. They know what usually happens next.

We do it differently. Not because we’re nicer. Because the strategy doesn’t work if the people who have to execute it don’t trust the process.

Why psychological safety is a strategic requirement

The best AI strategy in the world fails if the people responsible for implementing it are afraid. Afraid of looking incompetent. Afraid of losing their jobs. Afraid of being honest about what’s broken.

Fear produces three behaviors that kill every engagement:

Information hiding. The people who know the most about your systems are the ones closest to the problems. If they think honesty will be punished, they’ll tell you what you want to hear instead of what you need to know. Your discovery assessment will be built on a polished version of reality, and your strategy will be optimized for a system that doesn’t exist.

Passive resistance. People who feel threatened don’t fight openly. They comply on the surface and drag their feet underneath. Meetings get rescheduled. Data requests take three weeks. Requirements change after every review. The project doesn’t fail. It just never finishes.

Talent flight. Your most capable people have options. If they sense the engagement is designed to replace them, they start interviewing. By the time the roadmap is ready, the people you need to execute it are gone.

None of these are character flaws. They’re survival behaviors. And they’re completely predictable.

How we handle it

Week one: the ground rules

Before we touch a single system, we have a conversation with every stakeholder group. Not about technology. About what this engagement is and what it isn’t.

We say it plainly: we are not here to find people to blame. We are not here to recommend layoffs. We are not here to prove that your current systems are bad. We are here to find out what’s costing your organization time and money, and to design a path forward that your team can execute.

We say this to the executive sponsor. We say it to the department heads. We say it to the people who do the work. We say the same thing to everyone, because trust fractures the moment people get different messages.

During discovery: honesty without exposure

When we interview your team, we’re asking them to be honest about problems they may have been hiding for years. The spreadsheet that runs a critical process. The workaround that exists because a system request was denied three years ago. The data that nobody trusts but everyone uses.

We protect the people who tell us the truth. Our assessment documents describe systems and processes, not individuals. “The reconciliation process requires 14 manual steps” does not become “Sarah has been doing this wrong for 10 years.” The problem is the process. The fix is the process. The person who told us about it is not named.

This is not a courtesy. It’s a methodology. If people learn that honesty in one engagement leads to exposure, they will never be honest with us again. Our reputation depends on being safe to talk to.

In the roadmap: roles evolve, they don’t evaporate

AI changes what people do. It rarely eliminates people entirely. The person who spends 60% of their time on manual reconciliation doesn’t lose their job. They get 60% of their time back for the work that actually requires judgment.

We design the roadmap to show where roles change, what new skills are needed, and what the transition looks like. Not as a footnote. As a section that gets the same weight as the technical architecture.

When we present the roadmap, the people affected should see themselves in the future state. If they can’t, we didn’t do our job.

What this looks like in practice

We worked with an organization where the IT team had been manually managing a process that should have been automated five years ago. They knew it. They were embarrassed about it. They’d been telling leadership “it’s fine” because asking for help felt like admitting failure.

In the first interview, the lead engineer said: “I need to know this won’t be used against us.”

We told him it wouldn’t. And then we proved it. The assessment described the process gap as a systemic issue caused by underfunding and shifting priorities, not as a team failure. The roadmap included automation for the manual process and training for the team on the new system. The engineer who was most afraid became the most engaged champion of the implementation.

That’s not a feel-good story. That’s the strategy working because the people inside it felt safe enough to participate.

The cost of skipping this

When consultants treat psychological safety as a soft skill or an HR concern, they build strategies on incomplete information. The discovery is shallow because people hid the real problems. The roadmap is rejected because the people who have to execute it weren’t brought along. The implementation stalls because nobody trusts the process.

We’ve seen it happen. We’ve been called in to fix it after it happened. It’s always more expensive the second time.

Our commitment

Every engagement starts with the same promise: we protect the people who tell us the truth. We design the future state with the current team in mind. We present findings in a way that describes systems, not individuals.

This is not optional. It’s not something we do when the client asks for it. It’s built into how we work because the work doesn’t produce results without it.

If you’re evaluating firms and want to understand what else to look for, start here: How to Evaluate an AI Strategy Firm (Including Us)

Bryant Herrman is the founder of Merivant, an AI-native strategy firm based in Los Angeles.