This month I completed the Certified Public Manager program through the State of New Hampshire, and the capstone project my team built was recognized with the NH Director's Award, selected by the NH Director of Personnel for providing substantial value toward improved operations of state government. Here's what that year taught me, and why it matters for your business.
I've spent most of my career at the intersection of business and technology. I know the tools. I understand the landscape. But spending a year embedded in a real AI adoption project inside New Hampshire state government taught me something I hadn't fully absorbed before: the technology is almost never the hard part.
The hard part is people.
What We Set Out to Do
Our capstone project was built around a straightforward but ambitious goal: design and deliver a scalable AI training program for New Hampshire's public workforce, starting with the Division of Children, Youth, and Families.
Before we wrote a single training module, we ran a survey. The numbers stopped us cold.
Sixty-eight percent of DCYF staff reported little to no experience with AI tools. Seventy-two percent said they wanted to learn. That gap, between where people were and where they wanted to be, was the real problem we were solving. Not a technology gap. A confidence and knowledge gap.
We also asked where AI could actually help. The top three answers: automating repetitive tasks (61%), improving communication workflows (54%), and enhancing data analysis (49%). That told us exactly where to aim the training, toward things people were already doing, not abstract AI theory.
What Happened When We Met People Where They Were
The results were measurable and, frankly, striking.
Before training, about 10% of participants described themselves as "very confident" in their understanding of AI. After training, that number jumped to 74%. The share of employees who reported using AI tools at least occasionally went from 39% to 79%. More than 80% said they believed AI would improve their efficiency and their decision-making.
Those aren't vanity metrics. That's a workforce that went from uncertain and skeptical to capable and engaged, in a matter of weeks, because the training was built around their actual jobs and their actual concerns rather than the technology itself.
The NH Director of Personnel recognized the project with the Director's Award, citing its potential to replicate across state agencies and its alignment with the Certified Public Manager program's goals: fostering innovation, reducing inefficiencies, and strengthening leadership capacity.
The Lesson That Transfers Directly to Your Business
Here's where I want to speak directly to business owners, because this is the part that gets missed in almost every AI conversation I hear.
Most small and mid-size business AI adoption fails not because the technology is hard to use or too expensive. It fails because leadership skips the human layer entirely. They buy a tool, hand it to the team with minimal guidance, and wonder why adoption is low or inconsistent. Sometimes they mandate it and get surface-level compliance without real engagement. Neither approach works.
What works is exactly what the government project proved: start with the people.
Three Things Your People Are Telling You, If You Listen
When we surveyed that workforce before training, the patterns we found aren't unique to government. They show up in every organization, including small businesses and nonprofits. Here's what your team is likely thinking right now, whether they're saying it out loud or not.
They're uncertain, not resistant. The 68% of DCYF staff with little to no AI experience weren't refusing to learn. They just hadn't been given a clear on-ramp. Most people don't resist new tools out of stubbornness. They resist because uncertainty feels uncomfortable and no one has made it safe to ask basic questions. The antidote is structured, low-pressure exposure before expectation.
They know where the friction is. Your team members are the people doing the actual work. They know exactly which tasks are repetitive, which handoffs are slow, and which communication loops waste time. Our survey found that DCYF employees pointed directly at automating repetitive tasks as their top AI priority. Your team can do the same. Ask them. Their answers will tell you where AI investment will actually land.
They want to succeed with it. Seventy-two percent of the staff we surveyed said they wanted to learn. That's a remarkable number in any organization, and it wasn't unique to government. People generally want to be good at their jobs. When AI is framed as a tool that helps them do their jobs better rather than as a threat to their position, the response shifts dramatically. Framing matters more than you'd think.
What a People-First AI Rollout Actually Looks Like
The approach we used in government translates directly to a business of 5 employees or 50. The principles are the same.
Start with a real assessment, not assumptions. Don't guess where your team is with AI. Ask them directly: what do they know, what worries them, and where do they spend the most time on tasks that feel repetitive or manual? That data shapes everything that follows. Without it, you're designing training for a team that exists only in your head.
Identify priority use cases from the work itself. Not all AI applications are equal in your context. The question isn't "what can AI do?" It's "what does AI do that would actually make a difference in how my team operates?" The answers are almost always specific: drafting client proposals, summarizing long documents, generating first drafts of routine communications, pulling patterns from data. Start there, not with the broadest possible scope.
Build in ethical guardrails from the start. One of the things we emphasized in the DCYF project was ethical, accessible, and job-relevant AI use. That's not a compliance checkbox. It's a framework that helps your team understand what AI is appropriate for and what it isn't, what to review carefully and what to trust more readily. Without that framework, you get inconsistent use and, eventually, mistakes that damage trust internally and externally.
Measure the shift, not just the adoption. Counting tool logins is the wrong metric. The right metrics are confidence, actual task behavior, and outcomes. Did the work get done faster? Did communication quality improve? Did your team's relationship with the tool go from hesitant to habitual? Those are the questions worth asking at 30, 60, and 90 days post-rollout.
Design for sustainability, not one-time training. A single training session doesn't stick. What sticks is ongoing access to resources, a team culture where questions are welcome, and leadership that models the behavior. In government, the goal was a replicable model across agencies. For a business, it's the same idea: build something that grows with the team rather than requiring constant restarts.
The ROI Is in the Confidence, Not Just the Capability
There's a number from our project that I keep coming back to: 10% to 74% confident.
That's not just a training outcome. That's 64 percentage points of people who went from feeling like AI was something happening to them to feeling like it was something they could use. The behavior change followed naturally. When people feel capable, they engage. When they engage, they figure out where the tool actually helps. When they figure out where it helps, the efficiency gains are real.
The businesses that will get the most out of AI investment over the next few years aren't the ones that deploy the most sophisticated tools. They're the ones that bring their people along intentionally, starting with curiosity and ending with capability.
Government taught me that. It's a lesson I'll carry into every client engagement I take on.
If you're not sure where your team actually stands with AI, or where the highest-value use cases are in your business, the AI Readiness Audit from Wilson Digital Strategy gives you that picture. It's a self-contained assessment you can work through independently, with a consulting conversation available if you want to go deeper.
