Sunday, July 19, 2026

Pegasystems Shifts AI Pricing Model Away From Token-Based Costs

June 10, 2026

Hyderabad — Pegasystems Inc., an enterprise AI software provider specializing in mission-critical operations, unveiled a redesigned approach to building and managing agentic workflows that dispenses with per-token pricing models. Announced at PegaWorld, the initiative allows organizations to construct and deploy AI agents across Pega Infinity 26 without incurring charges based on token consumption.

The company's strategy centers on the Pega Predictable AI architecture, which concentrates computational reasoning during the design phase rather than at runtime. This shift enables deployed agents to operate with greater speed, consistency, and significantly reduced operational expenses—directly tackling two major hurdles enterprises face when expanding AI agent deployments: mounting token expenses and inconsistent performance.

The Challenge of Scaling AI Agents

Enterprise leaders are confronting mounting expenses as AI initiatives transition from experimental phases into production environments. Large language model vendors have shifted away from flat-rate subscription models toward consumption-based pricing tied to token usage. This transition has caught many organizations off guard, particularly as complex requests demand additional reasoning steps, frequently resulting in unreliable and inconsistent outputs.

Architectural Approach

Pegasystems applies intensive AI reasoning during the design phase, leveraging artificial intelligence's problem-solving capabilities to reimagine legacy business processes and systems. Through Blueprint AI and the newly introduced Infinity Studio, design-focused agents assist teams in constructing optimal agentic workflows for critical business operations—such as managing customer service requests, processing loan approvals, handling claims underwriting, and improving patient care experiences.

Following workflow design and deployment, the platform transitions to a streamlined semantic AI mode for runtime operations. Rather than re-evaluating each new workflow, agents employ a lightweight AI query to interpret user intent, identify the appropriate Pega workflow, and execute it methodically. When specific workflow steps require deeper language model capabilities—such as document parsing or interaction summarization—the system supplies targeted, specific instructions to maintain predictability.

This dual-phase methodology produces two significant advantages. First, outcomes remain consistent and predictable, as agents follow pre-approved workflows rather than re-reasoning each interaction—a critical requirement for regulated industries. Second, costs become predictable and substantially lower, as AI reasoning occurs once during design rather than repeatedly during runtime, enabling cost-effective operation at scale.

Quantifying Potential Savings

Pegasystems developed an interactive AI Token Cost Calculator to help enterprises measure the financial benefits of this model. The tool projects potential savings by comparing Pegasystems' approach with token-metered alternatives, factoring in workflow volumes. Many clients may achieve savings exceeding 20 times their current expenditures, depending on workflow complexity and operational scale.

Pricing and Availability

Rather than charging per token or per user seat, Pega's model charges per completed "case"—a task executed from beginning to conclusion. When an AI agent handles a customer order modification, that single completed interaction counts as one case. Beginning in the third quarter, Pega Infinity 26 clients will pay a single, flat rate per completed case, irrespective of the extent of AI processing occurring behind the scenes. This approach directly aligns expenses with measurable business value.

Industry Perspective

Alan Trefler, founder and chief executive of Pega, remarked that enterprises are recognizing the unsustainability of token-focused billing models, which generate escalating costs and unpredictable outcomes. "AI generates maximum value through dependable outcomes delivered at scale," he stated. "We charge clients based on meaningful work accomplished rather than token consumption, and our architecture ensures governed execution, providing organizations with unprecedented flexibility in deploying AI agents."

Liz Miller, vice president and principal analyst at Constellation Research, noted that the industry has reached a phase where measurable value and accountability have become central considerations. "Business outcomes have frequently been conflated with AI outputs, leading some to uncritically embrace token-based metrics as indicators of agent effectiveness," she explained. "Vendors and organizations must now demonstrate concrete returns on investment, utilization levels, and justify rising expenses. Solutions that deploy tokens with high efficiency and purpose will furnish enterprises with a competitive edge, enabling continued expansion of their agentic capabilities."

Compare options