Pegasystems Infinity 26 Brings Predictable Pricing to Agentic CX

What happened

Pegasystems has unveiled its Infinity 26 update for Pega Customer Service, repositioning the platform as a governed orchestration layer that sits beneath the contact centre rather than competing head-on with full CCaaS suites. The release adds Model Context Protocol (MCP) connectivity, new agentic service capabilities, and a pricing model the vendor calls Predictable AI, while leaving core telephony and channel infrastructure to partner ecosystems.

According to CX Today, the roadmap responds directly to a recurring complaint from Pega's customer base: AI capability alone means little if organisations cannot govern, audit or reliably budget for how autonomous agents behave once deployed. Infinity 26 is framed as an attempt to make agentic workflows both connectable — via MCP, an emerging standard for linking AI models to enterprise systems and data — and commercially predictable, addressing concerns about unpredictable consumption-based AI costs.

Why it matters

The update signals a broader shift in how enterprise software vendors are packaging agentic AI: not as a standalone chatbot feature, but as a workflow and governance layer that has to interoperate with whatever CCaaS, CRM or data infrastructure a customer already runs. By adopting MCP, Pega is betting that interoperability — rather than a closed, all-in-one platform — will be the deciding factor for enterprises choosing where to place their AI investment.

The Predictable AI pricing approach also speaks to a maturing buyer conversation. Early enterprise AI deployments have been dogged by usage-based pricing that is difficult to forecast at scale, and vendors that can offer clearer cost structures may have an edge as procurement teams move from pilots to production. For transformation leaders, the message is that agentic AI's next competitive battleground is less about model sophistication and more about governance, integration and commercial transparency.

The Renascence take

Vendors have spent two years selling the promise of agentic AI on capability alone; Pega's move suggests the market has quietly shifted the question from "what can it do" to "can we trust, audit and budget for it."

The real signal here isn't the feature list — it's that a major vendor is now selling governance and cost predictability as the headline benefit, ahead of raw AI capability. That's a tacit admission that most enterprises' agentic AI pilots have stalled not on performance but on control: nobody wants an autonomous agent making service decisions they can't explain to a regulator or a CFO. Operators evaluating agentic service platforms should stop scoring vendors on demo impressiveness and start scoring them on how clearly they can answer three questions: what does this cost at scale, who can audit an agent's decision after the fact, and how does it plug into the systems we already have. The vendors that answer those plainly, rather than defaulting to "it depends," are the ones actually ready for production.