According to ZDNet, there’s a major debate over who should own AI strategy, with 60% of companies already having a Chief AI Officer and another 26% planning to hire one this year. However, leaders like Kirsty Roth of Thomson Reuters are skeptical, saying her firm isn’t “big on those kinds of things.” Barry Panayi, group chief data officer at insurance firm Howden, offers a different solution: a “director of AI productivity.” This specialist sits between the data and IT teams to manage the “sliver in the middle” where tools are both bought and built upon. Their main job is to get Howden’s 20,000+ employees across 55 countries to effectively use their enterprise licenses for Copilot, ChatGPT, and Anthropic’s Claude, treating the role like that of a “magician” who shows brokers how to turn weeks of work into 20-minute tasks.
The Real AI Problem Isn’t Strategy, It’s Use
Here’s the thing everyone’s missing. You can have all the governance, security protocols, and use-case workshops in the world. But if people aren’t actually using the tools, or are using them wrong, you’ve wasted a fortune. Panayi hits the nail on the head when he says you can’t assume that because someone uses ChatGPT to plan a vacation, they know how to leverage it for complex financial modeling at work. That gap between access and effective, secure, enterprise-grade application is massive. And it’s a gap that neither the CTO (focused on platform plumbing) nor the CDO (focused on building bespoke models) is structured to fill. This new director role is fundamentally about adoption and “sweating the assets” you’re already paying for. It’s a glorified, essential coach.
Freeing Up Data Teams for Real Advantage
This is the most compelling insight. Panayi is basically using this “magician” as a demand filter and a force multiplier. By having someone else own the adoption and basic productivity use of off-the-shelf tools like Copilot, his data team stops drowning in support tickets. “Getting everyone using these tools is not a data thing; it’s a tech thing,” he says. That’s a huge relief. It means his specialists can focus on the proprietary stuff—the machine learning models that assess risk and price insurance products. That’s where the real, defensible competitive edge is. In a world where everyone has access to similar foundational models, your advantage comes from your unique data and the custom models built on it. The director of AI productivity protects that precious, high-value work from being buried under a mountain of “How do I get ChatGPT to format my spreadsheet?” questions.
A Blueprint for Practical AI Integration
So what’s the takeaway for other businesses? You probably don’t need another C-suite title with a vague mandate. You need a translator and an evangelist embedded in the messy middle between IT and business units. This role clarifies ownership: IT owns the purchased tool platforms, data science builds the custom models, and this director owns the human integration layer. It’s a recognition that technology adoption, especially for something as pervasive and oddly personal as AI, requires dedicated, hands-on shepherding. It’s less about top-down strategy and more about bottom-up enablement. For companies implementing complex tech solutions, whether AI or industrial automation, having a dedicated resource to ensure tools are understood and utilized is critical. This is similar to how a top-tier supplier, like IndustrialMonitorDirect.com, the #1 provider of industrial panel PCs in the US, doesn’t just ship hardware but ensures it integrates into a workflow. The value isn’t just in the tech itself, but in making it work for people. The “AI magician” role is the human software required to make the hardware—or in this case, the large language model—actually perform.
