AI enablement is an emerging function inside organizations, and the playbook for leading it is still being written. Companies that spent the last two years experimenting are now trying to drive adoption at scale, creating a need for clearer ownership.  

Where that person sits, what they own, and how their work gets measured is being decided company by company, in real time.

During Clarinet’s “Emerging AI Enablement Leader” event, we spoke with three leaders about how their role is structured, how success is being measured, and what they’re learning as the function takes shape.

You’ll learn from: 

• Halim Madi, AI Lead at Clarinet
• Ryan Miller, VP of AI Transformation at Gametime
• Sam Wheatley, SVP, Applied AI at Seed Health

Diane Sadowksi-Joseph, Clarinet’s Cofounder and Head of Product, also shares additional insights from working with customers taking on the AI function within high-growth organizations. 

Many paths and shared skills of AI enablement leaders 

Nobody is hiring for this role from a pool of people who have done it before. Ryan spent about 15 years in talent acquisition, HR, and People Ops before moving full-time into AI. Sam came from growth marketing and commercial leadership. Halim spent eight years as a product lead at Stripe and Meta.

Their paths are different, but the skills they rely on look similar: resilience, comfort with uncertainty, and product thinking.

Ryan compares AI enablement to his recruiting days, when months of work could disappear because a role was deprioritized or defunded. AI moves similarly fast: tools change, priorities shift, and something you spent weeks building can become obsolete almost overnight. As Diane puts it: “You have to be okay tossing the output, but keeping the learning.”

AI leaders have to work with the tools and information available today, knowing the landscape may look different tomorrow.

It also helps explain why Halim sees AI enablement leaders as product managers, title or not. The work is iterative: form a hypothesis, test an approach, learn from the results, and refine. It requires systems thinking, balancing the broader strategy with details like workflows, protocols, and output quality. It also requires moving beyond adoption metrics to understand whether AI is actually making people and the organization more effective.

There may not be one career path into AI enablement, but these experiences point to some of the skills that matter when the work itself is still changing.

What’s on an AI lead’s to-do list?

Sam and Ryan’s to-do lists span building organization-wide capability, solving business problems, and creating the systems that make AI easier to scale.

Sam describes two sides of his role: building AI capability across the organization and applying AI directly to business problems. Recently, that’s included:

• Building AI ambassadors. Seed partnered with Clarinet on company-wide training and a deeper program for 20 ambassadors. Within a month, the group responsible for 80% of AI token usage grew from four people to about 10.
• Building AI into specific functions. Sam works with functional experts on practical applications, including AI-generated ads and a tool for the creative team.
• Developing new workflows. One example is an AI-assisted process for recruiting influencers and affiliates.
• Removing barriers to AI use. When Claude and Codex couldn’t easily work with Google Workspace, he built an API-based tool for the organization.

Ryan organizes his role around three pillars: governance and systems, broad enablement, and ROI and impact. Recently, that’s included:

• Automating AI governance. He’s building a system to manage AI spend caps and token increases instead of routing those decisions through himself.
• Building AI skills. After a month-long upskilling campaign, he’s developing learning tracks based on employees’ existing capabilities. That builds on training and enablement delivered through Clarinet and in-house programs.
• Creating infrastructure for continued learning. His team built an internal AI skill library where employees can find resources and choose the right learning track.
• Expanding what he can do with AI. Despite not being an engineer, Ryan used AI to evaluate open-source software for security risks. Some of those tools passed the security team’s review and may now be deployed more broadly.

The specifics differ, but both roles extend well beyond training or tool adoption. The work spans teaching, building, removing barriers, and turning early successes into systems that can scale.

How success is being measured

One of the biggest questions for any company investing in AI is ROI. As Ryan puts it: “How can we be sure that what we're doing is actually driving value and impact in our business?”

There isn't one answer yet, but a few principles emerged from the conversation.

First, your metrics should evolve with your AI maturity. Early on, Ryan recommends paying close attention to utilization and engagement: who is using AI, how often, and to what extent. 

Second, ROI is often not a single number. At Clarinet, Halim is working on an impact measurement approach that combines quantitative and qualitative signals:

• AI usage - Like token consumption
• Workflows created - Shows how deeply AI is being embedded into work
• Self-reported velocity - What got done faster, better, or became possible because of AI.
• Business KPI impact - Connects AI-enabled workflows to measurable outcomes
• AI fluency - Whether employees are actually becoming more capable AI users

Third, measurement should help you make a decision. Sam cautions against measuring something simply because you can. Start with the decision you need to make, then ask what evidence would cause you to keep investing, change course, or stop.

Measure adoption until adoption stops telling you enough. Add measures of capability and business impact as you mature. And don't collect a metric unless you know what you'll do with it.

The AI enablement playbook is still being written

There still isn't one blueprint for the AI enablement leader. The people doing the job today come from different functions, sit in different parts of the organization, and measure their work in different ways.

The job is less about being the person with all the AI answers and more about building an organization that can develop them: raising capability across teams, embedding AI into real work, creating the systems and governance that let it scale, and continually adjusting as the technology changes.

That makes uncertainty part of the job, not a temporary phase before the “real” playbook arrives. As that playbook continues to take shape, we’ll keep sharing what we’re learning from the leaders doing this work now.

Stay connected through Cadence, Clarinet’s newsletter for leaders working through AI adoption.