How do you know whether AI is making your company smarter rather than simply filling dashboards with impressive activity?
In this episode of AI at Work, I speak with Michael Speranza, CEO of Kantata, about why familiar productivity metrics may be giving business leaders an incomplete picture of AI ROI. Companies can measure time saved, tasks completed, and documents generated, but those figures say little about whether AI is improving commercial decisions, creating revenue, or producing better client outcomes.
Michael introduces the idea of the expertise compounding rate. This measures how effectively a company captures, synthesizes, shares, and builds upon the knowledge created through its projects and people. For professional services firms, that knowledge can include client conversations, previous deliverables, staffing decisions, financial performance, project outcomes, and relationships between colleagues.
We discuss how AI can connect that information through a business specific knowledge graph. A team beginning a new project could identify similar work, locate colleagues with relevant experience, understand previous outcomes, and make better staffing or pricing decisions. Institutional knowledge that previously sat inside documents, meeting transcripts, or an employee’s memory can become available at the point of decision.
Michael also shares an example of a services company using AI to change its project economics. By reducing delivery costs, the firm could offer projects at prices that created a viable business case for clients who previously would have postponed the work. That suggests AI ROI could be measured through sales conversion, opportunity close times, revenue growth, and the ability to expand without adding headcount at the same rate.
Kantata frames the wider market around a revealing paradox. AI adoption across professional services reportedly increased by 40 percent last year, while executive confidence in real time visibility declined and revenue growth slowed to roughly half the industry’s historical benchmark. Greater adoption alone clearly does not guarantee stronger results.
Michael argues that efficiency has become the price of admission. The commercial advantage comes from making each project more informed, predictable, and valuable than the one before it. We consider what leaders should measure, how human expertise and AI resources may influence future pricing models, and why clients care far more about outcomes than invisible automation behind the scenes.
If every project created knowledge that improved the next one, how would that change the way your company measures AI ROI? Listen to the conversation and share your thoughts with me.
