Used AI to solve a database performance problem, without sharing PHI

navitend’s software team has worked alongside this organization’s leaders and staff for years to modernize their core application. Along the way we introduced authentication, authorization, and auditing, added role-based access control, and integrated the application with Microsoft Business Central to modernize its finance function.

As the application took on more, the database became the bottleneck. Users saw slowdowns and database deadlocks, and “it’s running slow today” became a familiar complaint. Finding the cause meant tracing the problem through a large, long-lived codebase and the data structure beneath it.

Rather than spend weeks tracing the problem by hand, navitend’s team used AI to analyze the application’s source code and its database design. The analysis pointed to specific programming and database design changes, which our developers reviewed, tested, and made.

We did it responsibly. The AI worked with the code and the structure of the database, never the records in it: no protected health information (PHI) was shared with any AI tool, and all of the work stayed within the bounds of the Business Associate Agreements (BAAs) already in place. As always, our developers decided what to ask and checked what came back.

Once the recommended changes were made, the database deadlocks were eliminated, and the application went from “it’s running slow today” to no user complaints.

AI helped the team find and fix a hard problem faster than tracing it by hand, while the organization’s compliance obligations stayed fully intact.