Pegasystems launches AI agents for finance sector control

Pegasystems has introduced a suite of AI agents aimed at providing greater control and transparency for regulated sectors such as financial services and insurance.

These new capabilities, called Pega Predictable AI Agents, are designed to address concerns about the use of agentic AI in industries where precision and accountability are critical to operations. Regulators and businesses have expressed reluctance about deploying AI agents in sensitive areas such as credit card dispute resolution and insurance claims assessment due to the potential for even minor inaccuracies to result in significant financial and reputational risks.

Pegasystems has built its Predictable AI Agents to be implemented alongside enterprise workflow automation, rather than as stand-alone "AI black boxes". The offering allows businesses to design, deploy, and audit AI-driven processes within a controlled environment, reducing the likelihood of unpredictable behaviour.

The agent suite includes several modules tailored for different functions within enterprise environments. Design Agents collaborate with human experts to create workflows and customer engagement strategies, relying on Pega Blueprint, which utilises multiple AI models to generate end-to-end solutions such as data structures, interfaces, security configurations, and user screens. When applications are generated using this approach, specific agents for additional cognitive tasks are also created.

Conversation Agents operate across various communication platforms, including chat, web, email, and voice. These agents work by linking directly to the underlying workflows, ensuring requests are actioned consistently rather than based on shifting prompt configurations. The company asserts this approach provides a more predictable experience than previously available methods.

Automation Agents are intended to handle routine tasks such as voice interpretation, summarisation of interactions, document analysis, and data extraction. These tasks are incorporated as defined steps within broader workflows, which the Pegasystems platform coordinates for transparency and reliability.

Knowledge Agents draw insights from both structured and unstructured data, integrating information from sources such as document repositories or external databases. The incorporation of Pega Knowledge Buddy agents is intended to embed industry best practices directly into operational workflows.

Coach Agents, such as Pega Coach, work with human employees to provide real-time, contextual guidance at particular points in the workflow. This is done to promote best practices and maintain consistency in complex or regulated processes.

Pegasystems' enterprise case management and access controls cover all predictable AI agents. A company spokesperson emphasised that this framework provides not just predictability but also heightened transparency, visibility, and security for user organisations.

The AI agents are available to use within Pega Blueprint and will be incorporated into the company's larger Pega Infinity platform in the third quarter of 2025.

"Enterprises have been hungry for a way to integrate AI that is predictable and transparent. Pega Predictable AI Agents helps ensure businesses can leverage agentic AI without sacrificing governance, transparency, or security. We've pulled back the curtain to deliver reasoning and workflow in a unified system that makes AI more predictable and manageable at scale," Kerim Akgonul, Chief Product Officer at Pegasystems, commented.

"Too many agent solutions are naïve and irresponsible 'AI black box' science projects. Pega is bringing much needed consistency and trust to AI agent deployments. Pega Blueprint has demonstrated the power of agents to reimagine the workflows that define a winning business. With Pega Predictable AI Agents, our clients can know that every agent is predictable, audited, and optimised for business success," Alan Trefler, Founder and Chief Executive Officer at Pegasystems, said.

Pegasystems' announcement arrives amid ongoing industry debate over the readiness and reliability of AI agents in regulated environments, where compliance and auditability are considered essential. The firm's focus on visibility and governance is intended to address longstanding reservations among organisations that require strict adherence to regulatory standards.