Navrina Singh

You know, most of our customers, Global 2000s, are building a lot of third-party AI, but they're also embedding a lot of AI within their own first-party applications.

In that case, you can imagine the way that they get started is, one, by registering these AI applications within Credo AI software.

So we have a very active AI registry, which can map out all your AI applications.

It can map out AI models that are powering those applications.

It can also basically demonstrate if you're building your first party applications, where those data sets are coming from.

And now we have a new capability around also logging your agents.

So within this registry, you can see a very comprehensive view of

everything from AI use cases to AI agents that you need to manage as an organization.

And that's literally the first step.

And once you've done that first step, I do want to give you a little bit insight into how Credo AI basically does governance.

From day one at Credo AI, we were extensively focused on industry-specific, use case-specific governance.

As you can imagine, AI is not a monolithic one technology.

It really depends upon the context of use.

So Credo AI's platform basically governs at the use case level.

And so as you start, you know, after you've registered your AI applications, we take you through either the risk path or the compliance path to manage holistically what is the risk management of that AI application or the compliance management look like.

And, you know, Grant, like I think just to for the listeners of this podcast, a little bit grounding is necessary right now.

When a new model releases, the first thing you will see is a lot of excitement around how it is showing on different benchmarks, whether it is, you know, a sweep bench, whether it is MMLU, whether it is ARC or any other benchmark.

Now, that is really exciting.

However, that is not useful completely for a business.

And let me give you a very concrete example.