Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

3 ways leaders can encourage successful AI adoption

Access to AI tools is not the same as knowing how to use them.

[Photo: Getty Images]

Profile photo of Daniele Grassi

CEO · General Assembly

Every few months, another headline declares an AI revolution. And by the numbers, they’re not wrong—more than eight in 10 human resources professionals are now using AI as part of their day-to-day work, according to the latest General Assembly’s State of Tech Talent report.

But when you ask a follow-up question, what are they actually doing with it? Everyone looks to someone else to answer.

Are employees using AI to draft emails a little faster and calling it a day? Or have they fundamentally redesigned how they approach their work? Most employers don’t know.

Underneath it all lies the bigger problem: Organizations are mistaking early AI use for real progress, and treating access and experimentation as a proxy for actual improvement.

“AI adoption” has become the corporate equivalent of a participation trophy. Companies proudly announce the new initiative, track it, report uptake to the board and investors, and in doing so, act as though access to a tool is the same thing as knowing when to use it.

For many, the phrase now describes tool deployment rather than a change in workforce capability. AI platforms have been rolled out across organizations, and introductory training programs are now almost standard. But access alone doesn’t meaningfully change how work is performed, and AI pilots by definition don’t scale on their own.

The gap is already becoming visible. In another recent General Assembly survey, 88% of leaders said their workforce was prepared for their AI ambitions, yet 61% had abandoned at least one AI initiative in the past year due to a lack of internal skills.

3 ways leaders can achieve successful AI adoption

1. Contextualize training according to workflows

Employees should develop the skills and habits for using new tools in ways that improve outcomes. So, not just understanding what AI can do, but also knowing when and how to use it in real work.

The gap between an employee who opens an AI application several times weekly to summarize meeting notes and one who has rebuilt their workflow around agentic tools (AI systems that can take actions and complete multi-step tasks independently) is a compounding productivity gap that’s as small as it will ever be.

And what starts as a gap between individual contributors eventually becomes a gap between teams and organizations.

Most AI training still looks like a generic course catalog: Long video modules, abstract use cases, and examples disconnected from specific roles, tools, and workflows. That kind of training can boost “hours completed,” but rarely changes how people design their days, structure projects, or make critical decisions.

Instead, effective AI training is embedded in real workflows. It teaches people how to redesign their day-to-day work using AI within their actual tools, data, and constraints, so the focus shifts from learning the tool to improving the work itself.

2. Set the standard

In our September 2025 survey, nearly three-quarters of leaders (73%) now encourage their teams to use AI, and 75% say their teams use AI tools regularly. Yet only 38% have restructured workflows around AI, and fewer than half (47%) report access to leadership-specific training.

Leaders who don’t use these tools can’t credibly set clear expectations for their teams, and they can’t easily distinguish better work from work that only looks more efficient.

When leaders experiment with AI in their own workflows—drafting materials, scraping data, pressure-testing decisions—they get a clearer view of what good use looks like and where potential risks lie.

From there, they can set clearer expectations for when AI should be used, what “better” actually means, and where it should be built into core workflows.

Leaders also need their own training to ensure they’re guiding their teams effectively. That makes it easier to ask teams to use AI where it will actually improve their work, instead of trying to keep up with the latest tools.

3. Track how work changes

When our team at General Assembly dug deeper, we found that 43% of human resources professionals say the difficulty of measuring the AI training’s value limits further investment in it.

Of course, it’s much easier to track who completed a module instead of tracking whether anyone changed how they worked the next day. Without that data, it’s hard to justify spending on training programs, let alone scaling them. And so pilot programs remain pilots.

Leaders can tell you how many licenses they’ve purchased or what percentage of employees completed a training pilot, but they often can’t answer a more basic question: Where has AI moved beyond experimentation and changed how work gets done?

While difficult to measure directly, organizations need to get comfortable using practical proxies (indirect measures that indicate whether work is changing).

A few examples of practical proxies:

  • Junior analysts take on work that used to require closer oversight
  • Teams move from draft to decision faster
  • Outputs improve without added time or headcount.
  • The problems employees feel confident tackling with AI after a training cycle
  • How often teams reuse AI-powered workflows instead of treating them as one-offs
  • Whether managers can point to specific projects where training changed the approach.

The real test of the AI revolution

We are probably living in the AI revolution. But the test is whether the work itself is changing and whether teams are making better decisions, moving faster, and producing higher-quality outcomes without adding time or headcount.

Leaders who contextualize training around their team’s work, take on AI tools themselves, and measure what changes turn AI adoption from a participation metric into a real advantage. They’ll be among the first to see that advantage show up in the work itself.

Daniele Grassi is president and CEO of General Assembly.