AI continues to show up in more and more of our day-to-day work, and what I like about this round’s examples is that they’re moving beyond experimentation and into real business value.

Hot off the presses, on the test-authoring side, Alex Y built a planning agent that can take Bluetooth specification evidence and turn it into requirements, proposed test designs, traceability, and a reviewer package. The important part is that it doesn’t guess. When required evidence is missing, it stops and flags the gap for human review.

In Engineering, the team continues to expand how they use AI during development. GitHub Copilot code reviews and cloud-hosted agent sessions are helping accelerate fixes, and they’ve also built dedicated security-review agents that can surface vulnerabilities before code is merged.

On the enablement side, Marketing and IT rolled out the Bluetooth Copilot presentation template, making it much easier for staff to create on-brand presentations without having to rebuild formatting, fonts, and layouts from scratch.

And Amber MCP continues to move forward, with validation work underway and pilot preparation in progress so Amber can be accessed directly from tools like Microsoft Copilot and other enterprise AI platforms. Amber MCP: Project planning established the path toward internal Copilot integration and an upcoming pilot.

The example I wanted to spend a little more time on is the Wolverines QA team.

As part of Project Sunset, the team had a dependency between the QualWorkspace automation suite and the legacy btCRM database. Using Claude Code Enterprise, they removed that dependency by consolidating test-account identity into a single shared JSON source. They then used AI to help validate the changes and confirm that no new test failures were introduced, followed by human verification of both the changes and the results. This was a Wolverines QA team accomplishment and a great example of applying AI to solve a real engineering problem. Actually, this slide doesn’t do this justice that AI could provide this accomplishment. Next slide, please.

What I really like about this example is the mindset: trust, but verify. AI helped accelerate the work, but the team still owned the outcome and verified the results. That’s exactly the kind of practical, responsible use of AI we’re trying to encourage across the organization.