How to Lead AI Adoption at Work—Before It Leads You

“Employees are ready for AI. The biggest barrier to success is leadership.”

That’s the headline from a recent McKinsey survey of 3,500+ employees and 230+ C-level execs across the US, Australia, India, New Zealand, Singapore, and the UK outlined in their report – Superagency in the workplace: Empowering people to unlock AI’s full potential.

The survey found that:

  • Employees are 3x more likely than leaders realise to already be using AI for 30% or more of their daily work.
  • Yet, C-suite leaders in the survey were 2.4x more likely to say “employee readiness” was the barrier to AI success than to recognise their own leadership as the issue.

In other words, while leaders are pointing at the team’s preparedness, the team’s already playing the game.

This mirrors what Cyberhaven found in their 2025 AI Adoption and Risk report:

“The majority of current AI usage falls under ‘shadow AI’—unsanctioned tools used predominantly by younger, mid-level employees and individual contributors, while adoption continues to lag among senior employees.”

So maybe the question isn’t: Are employees ready?

It’s: Are we ready to lead them?

When Innovation Moves Faster Than Strategy

I’m seeing this disconnect everywhere. While leaders wait for polished policies and enterprise-wide rollouts, teams are already charging ahead, solving problems with whatever tools help them move faster and think smarter.

The challenge is that, unlike past tech shifts, AI isn’t following a top-down playbook. It’s evolving from the ground up—informal, fast-moving, and full of potential and pitfalls.

That’s why the smartest leaders I’m working with have stopped asking, “How do we introduce AI to the business?” and started asking:

“Where is AI already in use—and how do we support it with the right safeguards, guidance, and leadership?”

What’s happening at the grassroots level is a huge opportunity. With the proper visibility, leadership and encouragement, those pockets of innovation become powerful momentum.

On the flip side: Without visibility or direction, even well-intentioned experimentation can open the door to real risks, like sensitive data being uploaded into public tools, or AI-generated content being shared with customers before it’s been sanity-checked.

Cyberhaven’s 2025 AI Adoption & Risk Report found that 34.8% of corporate data employees put into AI tools is sensitive—up from 27.4% just a year ago. Even more concerning, 83.8% of enterprise data going to AI is being directed to high-risk tools, rather than enterprise-grade, secure platforms.

The solution isn’t to stop the experimentation (which would kill innovation) or ignore the risks (which could be catastrophic). It’s to make the invisible visible and turn shadow innovation into supported innovation—and to do it sooner rather than later.

3 AI Leadership Strategies for Business Leaders

If you’re ready to lead what’s already in motion, start with these three moves:

1. Build a Living AI Strategy with Cross-Functional Experts

AI isn’t a ‘set and forget’ initiative—it’s a moving target. The tools are evolving fast, and so are the risks, so your strategy needs to keep pace.

Start by forming a small cross-functional working group—bring in IT, legal, risk, and a few early adopters already using AI on the ground. Together, map what’s happening, flag risks, and shape the right guardrails.

Ask the questions that matter: – What data should never be entered into AI tools? – Which platforms meet your bar? – How will you support safe, visible experimentation going forward?

Resources like Cyberhaven’s report serve as a helpful launchpad, outlining the 25 most commonly used AI tools and the 10 that pose the highest risk. For those ready to go deeper, KPMG’s Trusted AI Governance Roadmap and Secureframe’s practical guide (complete with a policy template) can help you build the guardrails your organisation needs.

Once you’ve started to shape your approach, don’t keep it under wraps.

2. Communicate Your AI Guardrails—Even if They’re Evolving

Gallup found that while nearly half of employees say their company is integrating AI, only 22% report a clear strategy, and just 30% indicate the presence of guardrails.

You don’t need perfect answers to lead well—you do need to keep your team in the loop.

Use 10 minutes at your next all-hands to share what you do know:

  • Which tools are safe
  • What data to avoid
  • Who to talk to with questions
  • What’s coming next

Gallup found employees are 3x more likely to feel prepared and 2.6x more confident using AI when their leaders communicate a clear plan.

3. Make AI Exploration Part of Team Culture

Right now, most AI experimentation is hidden. However, if you want to build capability, you must bring it into the light. That means making time, setting the tone, and showing your team that curiosity isn’t just welcome—it’s expected.

At Emergent, we’ve run two sets of AI sprints so far this year where our team have been encouraged to step out of business as usual and experiment with new use cases and AI platforms, then come and share back their progress and learnings to the team.

Canva recently paused “business as usual across every team, time zone and office” for a full week to dedicate time to AI learning and exploration. Block (owner of Afterpay) has implemented a similar approach.

Try carving out time for teams to test, tinker, and explore—within the boundaries of your evolving AI strategy. Whether it’s a focused afternoon, a hack day, or a team sprint, what matters most is the message you send: curiosity is encouraged, and this is a shared journey.

But you can’t just say ‘have a go’ and hope for the best. Make experimentation part of the rhythm. Let people see you trying, learning, adjusting—and they’ll follow.

The Real AI Leadership Challenge: It’s Already Happening

AI is already being used across your organisation—often under the radar. The real leadership challenge isn’t introducing AI, it’s catching up to your people, creating clarity around its use, and shaping a culture where safe, visible experimentation is supported.

Is your team innovating in the shadows—or are you lighting the path?