The integration of AI technology into USAA’s systems over the last five years exemplifies the company’s shift towards AI-based technologies during their digital transformation journey. USAA is broadly known as an insurer for military veterans and their family members; they have made significant investments into 200 AI tools and products that span the deployment of AI in underwriting, handling insurance claims, and automating back-office functions and customer service operations. The company seems to prefer a do-it-yourself approach to AI due to their reliance on pragmatism and experimentation rather than hands-on integration.
AI at USAA is “…given an experimental approach that is driven by strategy and careful consideration of the company’s broader objectives,” noted Bajaj, USAA AI Chief Data and Analytics Officer.
As USAA strives to balance caution with agility, these various partnerships help USAA accelerate its deployment of generative and other advanced AI systems. USAA employs a mix of ChatGPT, Google Gemini’s models, AWS’s Bedrock and Meta’s Llama while also developing tailored in-house models for certain functions like image recognition for claim assessments.
Generative AI for Interfacing Systems and Workflow Automation
USAA has integrated generative AI technologies into its customer-facing and hinter operations to improve response times during customer interactions. The company has deployed an AI system to help front-office staff respond to member questions faster by placing an AI system on top of 20,000 documents. USAA also uses software plugins like Code Assist and EagleGPT to automate coding processes and assist with inter- and intra-organizational workflows of information-based tasks.
AI technologies also enhance efficiency in data management processes, with Watsonx Orchestrate from IBM being implemented for workflow management in accessing data and approvals within the organization. The ability to autonomously execute tasks is critical for future automation initiatives. USAA maintains a risk-averse governance policy, assessing each AI application for functionality, impact, and alignment with compliance frameworks.
The AI council, led by Bajaj, assesses the feasibility of using AI in various use cases, considering its maturity and potential value for the organization and its members. All AI use cases undergo intake evaluation, model risk validation, and continuous performance evaluation to foster reliance in regulatory environments like financial services and ensure the performance and trustworthiness of their AI systems.
A Unified Data Foundation
USAA’s AI capabilities are enabled by a unified data environment, which is a system whereby a company’s interactions from all business channels are centralized.
“Over the last three years, we have created our unified data environment,” said Bajaj. “We have essentially taken all of our enterprise data and made it accessible from one place in one analytic environment, and we’ve connected the dots.”
With this data foundation, the company is able to extract insights relating to customer activities from banking, insurance, call centers, and digital platforms, which strengthen the AI applications that become smarter and more personalized.
Looking Ahead: Scaling Agentic AI
With the 2024 timeline, USAA intends to first focus on expanding its capabilities with agentic AI. The rollout will begin with lower-risk internal applications before moving to more comprehensive deployments.
“Adopting agentic AI will be important to achieving end-to-end automation, which will really enhance our productivity and experience,” Bajaj pointed out.
Nonetheless, none of these efforts are simple. As Bajaj described it, “The real challenge is: How do we transform those business processes to fully use AI solutions without disrupting our business?”
To solve this problem, USAA keeps a dedicated lab environment for continuous testing and iteration. The company seeks to implement AI that improves the automated processes in place without endangering its core mission of serving military members and their dependents by containing testing within their walls.
In an industry where change is often resisted, USAA offers a better model: integrating oversight with innovation and allowing every AI application to justify itself.
