Sapiens launches AI foundation for insurance platforms

Sapiens has launched Sapiens Brain, an artificial intelligence foundation for the insurance industry that sits at the centre of its SapiensAIP platform.

The product is designed to embed Sapiens' insurance data, software logic and documentation into an AI layer for insurers. The aim is to reduce model drift and hallucinations by grounding outputs in insurance-specific information built up over more than four decades.

Sapiens Brain underpins SapiensAIP, which the company recently introduced as an AI platform for insurers. According to Sapiens, the foundation draws on millions of lines of code, product documentation and design artefacts to give AI models access to insurance terminology, processes and regulatory context.

Paul Wheeler, Chief Executive Officer of Sapiens, said the launch reflects pressure on insurers to modernise systems and processes.

"There's a unique pressure within our industry to modernise, and AI is the key to unlocking that," Wheeler said.

He added that the combination of the two products is intended to connect AI agents with the company's existing systems.

"To be a disruptive player in insurance, we have taken AI further than any of our competitors. The unique combination of SapiensAIP and Sapiens Brain allows agents built outside our core systems to behave as though they were born within them - grafted back in like tissue, not bolted on like a robotic limb," Wheeler said.

Knowledge graph

Sapiens describes Sapiens Brain as an interconnected memory system based on insurance-specific knowledge rather than on general internet-trained models alone. The structure is intended to help AI models reason within the rules, exceptions and product structures used by insurers.

One example, according to the company, is Sapiens Brain for Sapiens Life, which draws on its history in life insurance products. Sapiens argued that this depth of sector-specific information would take a new entrant years to assemble.

"It's not an accident that we call it Sapiens Brain," said Glen Ceniza, Chief Technology Officer of Sapiens.

"A human brain is a system of interconnected memory, and that's exactly how we've engineered our product. AI models today are extraordinary, but a model by itself doesn't understand insurance - what it knows, it's learned from the open internet, with no ontology to reason inside. Without this grounding, a model will drift and hallucinate, however capable it is. Sapiens Brain gives the model the insurance-specific memory that's grounded, cited and hallucination-resistant by construction, enabling it to reason accurately, consistently and with explainability, which is exactly the standard this industry demands," Ceniza said.

Outside core systems

Sapiens said it chose to build SapiensAIP and the agents running on Sapiens Brain outside its core insurance platforms. That approach, it said, reflects the different development cycles of core software, which is built for stability, and AI systems, which change far more rapidly.

According to Sapiens, its core systems are exposed through Model Context Protocol, or MCP, allowing AI agents to connect with the core suite without the manual mapping commonly associated with traditional APIs. The architecture enables it to introduce AI updates more quickly than the release cycle typically associated with core software, the company said.

"SapiensAIP gives us access to value at a pace that's relevant for AI. We're not just making existing processes faster; we're rethinking what's possible in claims, underwriting, migration and configuration, which gives our customers the certainty and clarity they've never had before," Wheeler said.

Industry response

The launch comes as insurers test how generative and agent-based AI can be used in underwriting, claims, service and administration while managing regulatory scrutiny and operational risk. One of the central questions for the sector is whether large language models can be trusted in workflows that affect pricing, policy terms and claims outcomes.

Karlyn Carnahan, Head of Insurance, North America, at Celent, said insurers may gain more from redesigning work than from simply adding AI tools.

"Agentic AI won't simply make insurance processes faster - it will change how work gets done. As agents take on more of the coordination, information gathering and routine execution, human expertise can shift to the decisions where judgment creates the greatest value. The competitive advantage won't come from having the most AI; it will come from redesigning the business to use it well," Carnahan said.

MiWay, a South African insurer, also commented on the approach.

"The opportunity isn't just to automate insurance - it's to remove the parts that don't create value, so our people can focus on the parts that do. Used well, AI may make us more human, not less. At MiWay, we see Sapiens Brain as providing the intelligent foundation for many of our use cases," said Janes Du Toit, Chief Information Officer of South African insurance provider MiWay.

Trust and oversight

Sapiens said the system has been designed for regulated insurance environments, with audit trails and inspection records for AI decisions and reasoning chains. It is using domain-tuned models for specific tasks rather than relying on one model across an entire workflow, the company said.

"In this industry, AI is held to a higher standard than people are, and rightly so," Ceniza said.

"That's why every decision our agents make is logged and auditable, and why we deploy fit-for-purpose, domain-tuned models for contained tasks rather than one frontier model let loose across an entire workflow. Trust in AI comes from a quality standard - accurate decisioning, consistency and explainability - and that's the benchmark Sapiens is setting for the market," Ceniza said.