FICO Announces New Unparalleled Domain-Specific FICO Foundation Model for Financial Services Product Offering

FICO announced a new unparalleled domain-specific FICO Foundation Model for Financial Services (FICO FFM) product offering, consisting of FICO Focused Language Model for Financial Services (FICO FLM) and FICO Focused Sequence Model for Financial Services(FICO FSM), purpose-built to reduce hallucinations and achieve precise accuracy over conventional Generative AI (GenAI) models. The FICO FFM is designed with a precise, auditable focus on the data used to build and further task-train the model for each specific task or business problem. The model is architected for trusted, responsible use of GenAI.

Unlike general-purpose LLMs trained on broad world knowledge and requiring massive compute power, FICO?s focused domain-specific models require up to 1,000x fewer resources?making them significantly more focused, cost-effective to train and maintain, auditable and adaptable. FICO?s models are custom-built, engineered in-house and are focused on curated data sets related to the domain of financial services. FICO FLM is built on domain-specific and task-specific contextual data, allowing the model to significantly reduce hallucinations given that the training data is directly relevant to the problem.

These models are smaller, more accurate, and significantly lower in cost, driving a practical approach toward agentic AI. FLM models enhance performance, have a low barrier to entry, and streamline operations for financial institutions. FICO FSM is focused on leveraging long-range attention in transaction sequences, finding critical relationships in transaction histories typically not captured in traditional analytics systems. The model plays a crucial role in improving real-time detection accuracy across financial services transaction analytics, while capturing complex inter-relationship sequences in use cases like payment fraud, real-time risk assessment, next-best action, and more.

The model empowers customers to uncover transactional behavior patterns that were previously impossible or too expensive to identify. FICO FLM and FICO FSM are built with both patented and patent-pending Trust Scores, providing a risk-ranking of the reliability of the output generated by these models. By using a Trust Score, organizations can set their own risk thresholds, reduce hallucinations, and operationalize model outputs with confidence.

This framework empowers financial institutions to deploy GenAI responsibly, ensuring that all outputs align with business-defined knowledge anchors and have sufficient statistical support, enabling ongoing risk oversight and risk monitoring in GenAI decision-making. FICO has filed multiple patent applications on FICO FLM and FICO FSM core technology and specific use cases, including trust scoring frameworks, model training techniques, tracking content generated by language models, real-time monitoring, and transaction sequence modeling. These advancements underscore FICO?s innovation in Responsible AI and in the commitment to make GenAI usable, trustworthy, and impactful for the financial services industry.