Pegasystems Touts Governed AI Workflows, Token-Free Pricing at Oppenheimer Conference

Pegasystems is positioning its platform for regulated, control-heavy enterprises that need predictable and governed workflows, arguing that Pega can evolve more reliably than custom-built software. Its approach combines workflow rules with AI, human oversight and customer-selected agents or gateways. Pega said customers will not pay separately for AI tokens used within its platform. The company plans to manage costs by applying AI selectively, choosing appropriate models and using deterministic workflows where they are more effective. After weaker first-half annual contract value growth, Pega is strengthening its sales pipeline and pursuing new customers through increased outbound activity and revised incentives. Management remains focused on free-cash-flow margins above 30%, a “rule of 40-plus” profile and primarily organic growth.

Pegasystems NASDAQ: PEGA COO and CFO Ken Stillwell said the company is positioning its platform around enterprise workflows that require consistent, governed and predictable outcomes, particularly in regulated or control-heavy environments. Speaking at Oppenheimer’s 29th Annual Technology Conference, Stillwell described Pega’s core market as large organizations that need to configure specialized workflows rather than rely on off-the-shelf applications. He said the company has historically competed with internally developed software, arguing that custom code can create sustainability and change-management challenges over time.

Stillwell said Pega has incorporated artificial intelligence into application design, development, maintenance and workflow execution. The company’s approach allows customers to use Pega’s AI capabilities, their own agents, or their own gateways within workflows, he said.

He emphasized that generative AI alone is not suited to every enterprise task. Using bank loan origination as an example, Stillwell said the process can involve credit ratings, appraisals, underwriting, disclosure requirements and fair-lending rules. Those workflows must be applied consistently in order for banks to demonstrate compliance with regulations, he said. According to Stillwell, generative AI produces a unique response each time and therefore cannot by itself provide the deterministic outcomes needed for highly governed processes. He also cautioned that using AI agents to create a company’s own workflow systems could produce expanding and difficult-to-manage code bases. Stillwell said Pega uses agentic engineering in its own research and development work and has seen the need for careful human oversight. He said agents may attempt shortcuts when instructed to accomplish a task, potentially creating bugs or other unintended code behavior.

Stillwell also discussed the growing focus on AI computing costs, including token usage. He said organizations should consider both when AI is necessary and which model is most appropriate for a given task, rather than automatically relying on the most expensive frontier models. Pega’s commitment, he said, is that customers do not pay separately for tokens used within Pega. Instead, the company seeks to manage those costs int...