SS&C Blue Prism aims to guide companies in the transition from Robotic Process Automation (RPA) to automation based on AI agents, a necessary evolution to manage the increasing complexity of modern workflows.

The complexity of modern workflows

According to Steven Colquitt, VP Software Engineering at SS&C Blue Prism, today's workflows present a level of complexity that exceeds the capabilities of traditional RPA. Unstructured data comes from heterogeneous sources and interactions in the real world are increasingly less deterministic. Inputs and outcomes can vary, and decisions depend on the context in real time.

Brian Halpin, Managing Director, Automation at SS&C Blue Prism, cites the example of a credit agreement, where it is necessary to extrapolate numerous answers, underlining the level of reasoning required by large language models (LLMs).

A gradual path towards agentic AI

Halpin highlights that the goal is to provide an AI agent with the desired outcome, without specifying the detailed instructions to achieve it. However, the market is not yet ready for fully autonomous automation due to issues related to trust, regulations, auditability, stability and security. LLMs are subject to hallucinations and variations over time, and changing the underlying model can alter responses.

SS&C Blue Prism intends to help its customers, who already use process automation, to integrate AI into their activities. In many companies, AI is managed by a separate unit, and the goal is to promote integration between process automation teams and AI capabilities, to achieve further improvement in efficiency.

New technologies for agentic AI

SS&C Blue Prism will soon launch new technologies to help companies develop and embed AI agents within workflows, facilitating orchestration. The company previewed these innovations at TechEx Global, during the Intelligent Automation conference.

SS&C Technologies is one of the largest users of RPA in the world, with over three and a half thousand digital workers deployed. The company estimates saving hundreds of millions thanks to these systems and has already implemented approximately 35 AI agents in production for complex tasks.