Driving AI Adoption: What CLOs Learned at CLO Exchange New York

Your company bought the AI licenses. So why isn’t anyone using them?

Most companies have solved AI access. AI adoption is a different story. At CLO Exchange New York, Yoodli CEO and co-founder Varun Puri and Mike Mather, Yoodli’s Chief Learning Advisor and former Chief Learning Officer at KPMG US, put a familiar problem in front of a room full of learning leaders. Imagine a company buys 10,000 Gemini licenses. Three months later, only 10% of employees use them. What do you do next?

For many in the room, this was not a hypothetical. In fact, it is one of the most common stories in enterprise AI right now. Budgets get approved, tools get rolled out, and yet daily usage stalls.

The session, “Driving AI Adoption: How to get your teams to actually use AI,” also marked Mike’s first official day at Yoodli. As a result, the conversation blended a CEO’s view of the market with a CLO’s view of what actually changes behavior inside a large organization. Here is what they shared.

The biggest mass upskilling the workforce has ever seen

Varun opened with a bold claim: AI will drive the largest mass upskilling effort in the history of the workforce. However, he was quick to point out that we have lived through parts of this story before.

After World War II, the GI Bill of 1944 helped 7.8 million veterans reskill for a new economy. Back then, machines replaced tasks, but skills decided who thrived. Today, the same pattern is repeating. AI is the next shift, and L&D is its GI Bill.

In other words, the technology alone does not determine the outcome. Instead, the organizations that invest in their people’s skills will be the ones that capture the value. As Mike put it, “This is L&D’s moment.”

Why AI adoption stalls with traditional training

So why do so many AI rollouts stall? To answer that, Varun and Mike walked through how most companies approach AI training for employees today.

Typically, organizations run webinars, seminars, prerecorded videos and quizzes. Then, they measure success through participation, CSAT and NPS, with only a loose connection to whether people actually use the tool afterward.

The problem is simple. This model measures completions, not comprehension. Someone can finish a one-hour webinar on prompting and still never open the tool again. Consequently, leaders end up with strong attendance numbers and weak adoption.

Varun made the same case earlier this year in Fast Company, in his op-ed Why your employees aren’t using the AI you bought. As he noted there, enterprises spent $37 billion on AI in 2025, yet only 7.5% of employees have received anything close to extensive AI training. In short, the gap is not motivation. People want to use AI; they just have not had a real chance to practice it.

How to drive AI adoption: from license to habit

To show a better path, the pair ran a live demo built around the Gemini scenario. Rather than another course, the approach focused on three shifts:

1. Learn the tool in your own context. Employees learn with their own materials and workflows, not generic examples.
2. Practice a real workflow with AI roleplay. People rehearse the actual tasks they will use AI for, and they get instant feedback along the way.
3. Certify skills and see who is actually using it. Leaders get measurable proficiency instead of attendance counts.

This is the thinking behind Yoodli, the AI Platform for Human Skills. Enterprise teams use it to onboard employees, train people on AI tools, run certifications and put a coach inside their workflow. Specifically, content is built from your own materials, practice happens through AI roleplays guided by an AI tutor, skills get applied in real work, and results are certified and tied to outcomes.

Moreover, the approach is tool-agnostic. Whether a company has standardized on Claude, Gemini, ChatGPT or Copilot, the same method applies. It also extends well beyond tool fluency to manager training, sales enablement, partner training, onboarding, media training, AI coaching and helping leaders talk openly about AI and jobs.

What AI upskilling looks like in practice

Of course, frameworks only matter if they hold up at scale. That is why the session spent time on real customer results.

Customer	Use case	Result
Google Cloud	Certification	15,000 reps enabled org-wide, at the organization that built Gemini
$5B+ cybersecurity company	Ramp time	29% uplift from skills diagnosis to overall improvement in reps’ ability to ask questions
Government of El Salvador	Scaled training	About 1 million minutes of practice time for doctors

Together, these examples show the range. On one end, a global tech company certifies its own sales force on its own AI product. On the other, a national government uses AI practice to train doctors at a scale no in-person program could match.

Why we still need humans

Equally important, the speakers were candid about where AI falls short. Three limits came up:

• Behavior change is hard. Many people are still used to the old ways of learning, so new tools need support to stick.
• People hold AI to a higher bar. When AI glitches, frustration runs higher than when a human makes the same mistake.
• AI is not for decision making. AI is great for training. Still, it should not be used alone for screening or evaluating people.

For that reason, Yoodli’s view is that AI should help people, not replace them. The goal is to give every employee a practice partner and coach, while leaders and managers stay in charge of the decisions that matter. For more on this, see our recap of Yoodli and Google Cloud at SEC Denver, which covers the honest pitfalls of AI roleplay.

What to look for in an AI adoption partner

Finally, Varun and Mike left CLOs with four questions to ask any vendor helping drive AI adoption:

1. Do they work with the ecosystem? Look at how they partner with training providers you already use.
2. Are they AI-first by design? The platform should be built on AI, not bolted on.
3. Do they offer enterprise-grade security? It needs to fit your privacy and compliance rules.
4. Do they have a long-term vision? You want a roadmap that grows with your AI strategy.

Join the Yoodli Learning Advisory Board

To close the session, Yoodli announced the launch of the Yoodli Learning Advisory Board, a group of learning leaders shaping how organizations upskill their people for an AI-driven workplace. If you lead learning and development and want a seat at the table, reach out to the Yoodli team to learn more about joining.

Frequently asked questions about AI adoption

Why is AI adoption low even after companies buy licenses?
Most AI training measures attendance rather than skill. Employees finish a webinar but never practice the tool in their real work, so usage stalls.

How can L&D teams increase AI adoption?
Start with training built from the company’s own materials. Next, let employees practice real workflows with AI roleplay. Then, certify skills and track who is actually using the tool.

What is AI fluency training?
AI fluency training helps employees use approved AI tools such as Claude, Gemini, ChatGPT or Copilot confidently in their daily work, with measurable proficiency instead of course completions.

Should AI be used to evaluate employees?
According to Varun and Mike, AI is excellent for training and practice. However, it should not be used alone for screening or evaluation decisions.

What is Yoodli?
Yoodli is the AI Platform for Human Skills. Enterprise teams use it to onboard employees, train people on AI tools, run certifications and put a coach inside their workflow.