From the 2026 Technology Finance Symposium – West
Chris Price Director at Baker Tilly moderated a discussion with Ray Rike, founder and CEO of Benchmarkit, and Stuart Langer, deputy CTO at Sage and a CPA, focused on moving artificial intelligence beyond experimentation and into investments that deliver measurable enterprise value. Rike brought a metrics-driven perspective on AI economics, while Langer examined the technology, governance and operating structures required to deploy agentic AI responsibly at scale.
The discussion began by distinguishing agentic AI from more conventional workflow automation. Langer explained that predictable processes with structured data and clearly prescribed steps may be handled more economically through traditional automation. Agents become more compelling when organizations face unstructured data and dynamic processes where the technology must adapt and make choices. Still, he cautioned that the ability to deploy an agent doesn’t necessarily mean it’s the right investment.
For CFOs evaluating those investments, Rike emphasized starting with a financial hypothesis. Proposed AI initiatives should identify the business metric expected to improve, the magnitude and timing of that improvement, and the anticipated payback. The speakers also encouraged leaders to look beyond easily quantified productivity gains. Customer satisfaction, consistency, quality, and other operational outcomes can materially affect the ultimate return.
Calculating cost requires similar discipline. The panel noted that model subscriptions and token consumption represent only part of an AI initiative’s economics. Data storage and management, infrastructure, human oversight, testing, validation, and governance can significantly change the business case. Rike cited research showing that only 11% of companies could forecast AI costs within 10% of budget, underscoring the importance of attributing spending to individual agents and use cases.


