Optimizing a Responsible AI Program at a Global Financial Software Company
Problem
A global financial software company, operating a responsible AI (RAI) program to ensure its developers adhered to established frameworks for AI model development, saw an opportunity to make the program work harder in practice. The program already spanned multiple departments—legal, security, product, and support—each contributing to the processes that developers, engineers, and technical program managers were expected to follow.
The systems were in place, but day-to-day execution exposed friction. Users lacked clear visibility into the RAI program, struggled to identify the right people and resources to reach out to, and encountered multiple sources of truth for key metrics. Cleanup and remediation efforts had accumulated alongside these gaps. Looking to identify and eliminate the blockers slowing the program down and to surface opportunities for both the development and program teams, the company brought in Excelerate.
Approach
Excelerate began with stakeholder input sessions across the departments involved in the RAI program, gathering perspectives from legal, security, product, and support on where friction existed in the current process. In parallel, Excelerate conducted artifact analysis to inventory the resources, documentation, and tools the company already had at its disposal.
Through this work, Excelerate identified an existing internal asset—Google Vids—that the company had access to but had not yet operationalized in support of the rai program. Excelerate built dashboards to consolidate program metrics into a single source of truth and developed a template document to standardize how teams engaged with the program going forward.
Results
Excelerate delivered a training document, core training videos, and resource materials integrated directly into the client's knowledge base. This equipped developers, engineers, and technical program managers with clear guidance on how to navigate the RAI program. Consolidated dashboards replaced the multiple sources of truth that had previously created confusion around program metrics.
Most notably, the work drove a 2x increase in bot resolution rate. The frameworks were already in place; what the program needed was a way to make them findable, navigable, and consistently applied. By bringing visibility and shared tooling to existing systems, Excelerate helped the client get more from an investment they had already made.