Krishna Patel
Case study1 min read

Building an AI Registry Before Anyone Asked for One

How I created shared visibility into scattered AI work across SE and PS, before a bigger system existed to require it

The situation

AI work was happening everywhere across the SE and PS organization, individually, in isolated pockets, with no shared visibility.

Teams were solving the same problems without knowing it, good ideas were staying siloed, and leadership had no full picture of what was actually being built. There was no governance, no committee, no registry, nothing.

What I did

Designing for visibility, not gatekeeping

I built the registry to explicitly be a visibility program, not an approval bottleneck. Teams don't need permission to start AI work, the point was to make work visible, reusable, and easier to support, not slow anyone down.

Same work, happening twice. Now visible to everyone.

Building the system that runs itself

I set the registry up in Confluence with Rovo, with Slack integration and lifecycle automation, submissions get flagged for review automatically, the committee gets notified when something new comes in, and pending items get weekly reminders without anyone manually chasing them down.

Standing up the review process, not just the intake form

I built the committee's actual operating workflow, a decision guide with clear categories, a weekly review cadence, and the automation rules underneath it so the process runs on its own instead of depending on someone remembering to check a queue.

Intake, review, reminder. The loop runs without chasing.

Getting ahead of a bigger system before it existed

This started as our own SE/PS effort, before Business Transformation was building a larger, company-wide intake system. Building this early meant our organization was already gathering the right information and had a working practice in place, positioned to lead rather than wait.

The result

A working registry with real governance behind it, decision criteria, automation, a committee cadence, that gives leadership a full picture of AI work across the org and lets teams find collaborators and build on each other's work instead of duplicating it from scratch.