In this Brandon Hall Group™ Excellence at Work Podcast, Michael Rochelle, Chief Strategy Officer and Principal Analyst, sits down with Peter Mulford, Chief AI Officer and Head of Artificial Intelligence and Innovation at BTS. Together, they explore where organizations truly stand on their AI journeys, why so many are leaving value on the table, and what it actually takes to drive meaningful, measurable ROI from AI, not just incremental productivity gains.
Key Discussion Points:
- The Execution Layer Is Where the Real Opportunity Lives: Most organizations are focused on the model layer (Anthropic, OpenAI, Microsoft) and the application layer (interfaces, data governance, tools like Claude or Copilot), but the highest-leverage opportunity lies in the execution layer: how people actually interact with AI, and what they do or don’t do with it. This is precisely where HR and talent professionals have their greatest impact, and where too little attention is currently being paid.
- Most Companies Are Still Reading by Candlelight: A powerful analogy runs through the conversation. For years, we’ve used technology like a candle to read in the dark. AI is electricity. And yet, many organizations are simply replacing the candle with a lamp, using AI to summarize emails and boost personal productivity. The real opportunity, the equivalent of unlocking the fundamental physics of electricity, lies in frontier use cases that most organizations haven’t yet imagined.
- The Right Kind of Top-Down Involvement: Not all AI leadership is created equal. Peter distinguishes between two approaches: command-and-control (dictating who uses what tools and how), which tends to kill innovation energy and authorship; and criteria clarity, where leaders define the KPIs and problem areas that matter, then create space for teams to experiment and discover solutions. The firms seeing the most remarkable results are firmly in the second camp.
- AI Maturity Determines the Right Metrics: Where a company sits on the AI maturity curve should directly shape the ROI it pursues. Early-stage organizations should focus on adoption metrics and learn from how people naturally interact with AI. As maturity grows, the ambition of KPIs should grow with it, moving from personal productivity wins to workflow transformation and ultimately to high-impact business outcomes.
- Choosing AI-Enabled Vendors: What’s Different (and What Isn’t): Evaluating an AI-powered vendor isn’t entirely unlike evaluating any other technology provider. You examine the problem you’re solving, the model behind the solution, the application layer on top of it, and whether it reduces pain points or augments your people’s capabilities. What is different: leaders need enough working knowledge of how AI functions to ask the right questions about model training, data integration, and evaluation methods, so they can distinguish genuine capability from vendor hype.
- Identity Shifts Are the Hidden ROI: One of the most striking observations from BTS’s work is what happens to people’s sense of professional identity when AI actually works in their hands. Rather than feeling threatened, people who engage with AI experientially, working through real business problems rather than generic prompting exercises, often arrive at a realization: their judgment, experience, and domain expertise have become more valuable, not less. Knowing which AI output to trust, which to interrogate, and which to discard is a rarer and harder-won capability than the models themselves.
- You Don’t Need a Six-Month Program, But You Can’t Skip the Foundation: BTS has seen significant, third-party verified results in as little as a single day. But the formula requires two non-negotiables: first, a genuine technical foundation that demystifies how AI works (skipping this creates either reckless champions or fearful skeptics); and second, hands-on experience with frontier use cases, not safe, low-impact exercises, but pushing the models to their outer limits. That’s what opens people up to seeing what’s truly possible.
- The Highest-ROI Use Case: Synthetic Customers and Rapid Prototyping: One of Peter’s standout examples involves an FMCG company that used AI to create synthetic customer personas and run rapid prototyping on new product and market launches. It illustrates the category of use case with the greatest potential: quickly and cheaply testing ideas that were previously neither quick nor cheap to test. The result was a reimagined core solution that drove market wins, reduced dependency on external agencies, and delivered outcomes that were faster, cheaper, and better.
The conversation makes clear that the future of AI ROI isn’t about the technology itself, it’s about people. The organizations that are winning aren’t the ones with the most sophisticated models or the flashiest tools. They’re the ones who have figured out how to shift people from skepticism to ownership, aim collective creativity at problems that actually matter, and create the conditions for genuine breakthroughs. As Peter puts it, the goal isn’t to keep the candle burning a little brighter — it’s to discover what electricity can really do.
