TransUnion Deploys Gemini-Powered AI Agent
TransUnion has launched an AI agent built on Google Cloud's Gemini models to speed up financial analytics. The agent is designed to orchestrate complex workflows, handling data extraction and scenario analysis to reduce cycle times for credit intelligence.
The new "AI Analytics Orchestrator Agent" is built upon TransUnion's OneTru platform, a cloud-based system launched in March 2024 to unify the company's data analytics, credit risk, and fraud detection solutions. The agent utilizes Google's Vertex AI platform and Gemini models to translate natural-language questions into governed, production-ready analytical workflows. This is designed to reduce complex analytic cycles from weeks or months down to hours or even minutes. A key feature for enterprise use is the agent's emphasis on auditable and explainable AI. It breaks down each user prompt step-by-step, maps the steps to code, and provides the reasoning back to the user in plain language. This transparency is crucial for regulated industries like finance and is integrated with TransUnion's semantic knowledge graph to understand relationships between credit concepts, ensuring governed data retrieval and explainability. The agent is currently being used by TransUnion's 800+ internal data scientists to codify decades of domain knowledge into automated processes. The company plans to roll it out to external customers through its TruIQ solutions suite in the coming months, with future expansions planned for marketing and fraud use cases throughout 2026. This move reflects a broader trend of leveraging generative AI in financial services to enhance risk management and decision-making. AI models can analyze vast and varied datasets, including non-traditional sources like utility payments or rental history, to assess the creditworthiness of individuals with limited credit histories. This capability can improve loan approval rates while potentially reducing default rates.
