Agentic AI continues to transform modern business, which leaves legacy companies to face a critical question: evolve or be left behind. Few understand this crossroads better than Alan Trefler, founder and CEO of Pegasystems, who joined Bloomberg’s Tech Disruptors podcast to share how established enterprises can thrive in an AI-first world.
Pegasystems, founded in 1983, has survived and adapted through six generations of technology shifts. Trefler describes its evolution as deliberate. “From our inception, we aimed to create a better way for business teams… and IT teams… to operate more effectively in large businesses,” he said. Now, with AI driving the sixth wave of innovation, Pega has integrated what it calls “predictable AI” directly into its workflow and decisioning platform.
AI Without the Chaos
Trefler is quick to distinguish Pega’s approach to AI from the rest of the industry, especially in how it avoids the pitfalls of unpredictable generative AI reasoning at runtime. “We want to use the reasoning of these latest, most sophisticated models at design time,” he explained. “Once it defines the workflows, somebody can look at it, an auditor can approve it… Those workflows are defined at design time.”
This predictability, Trefler argues, is especially crucial for regulated industries like banking, insurance, and government, Pega’s primary clientele. “You can build a slew of workflows really easily that will always execute precisely,” he said. That’s a far cry from competitors that rely on language models to reason and make decisions on the fly, a practice Trefler calls “a horrible, horrible way to do it.”
Rethinking Legacy Systems
One of Pega’s most transformative tools is Blueprint, a generative AI-powered design agent. “You can feed in documentation from those [legacy] applications… and then we use the AI to take our best practices… and rethink what a system like this could be,” Trefler said. The result is a dramatic acceleration of transformation initiatives: “Being able to do things in literally a couple of weeks that used to take them many, many months.”
Trefler recounted a conversation with a client from a large bank: “They talked about they had over 6,000 applications… and they wanted to get rid of 2,400 of them.” Traditional approaches would take years. AI-driven pathways would speeden up modernizing critical infrastructure.
Unlike competitors still exploring how to monetize generative AI, Pega apparently made a strategic shift years ago. “We didn’t want to charge by the person… So we began charging by the unit of work,” said Trefler. This means customers pay for results (each onboarding, each case) not for seats or compute usage. “Now that Agentics is here… we actually find that that way of charging still works perfectly.”
The Agentic Future
Pegasystems is also looking beyond workflow automation. Its Agentic Experience Fabric introduces a conversational interface for users to interact directly with workflows without prompt engineering or retraining. “Our customers don’t have to go back and start going to some prompt studio… Those existing workflows can run agentically,” Trefler said.
By layering a language model over existing business logic, the system allows users to trigger complex actions through natural language. “It pulls all those workflows together, hooks them in with a language model to help you find the right one, and then helps you talk to it,” he explained.
This deterministic approach offers a powerful counter-narrative to the probabilistic chaos often associated with AI. “Those steps, those decisions, those updates to the backend systems, those had better be not just reliable, but predictable,” Trefler emphasized.
Though Pegasystems has historically served some of the world’s largest enterprises, including governments and global banks, it sees AI as a gateway to broader markets. “Our partners will be operating at several times the speed and velocity… and that will really help propel us out of just the super high-end,” said Trefler.
Their pathway seems to be a methodical, systems-based approach to modernization. “Despite all the confusion and all the hype… being able to have predictable outcomes is going to be really important,” he concluded.
