Use Case - Bug diagnosis

Let me show you how you can use Count AI to help you root cause bugs and come up with some fixes. So here we have in the canvas an event stream of data, just a usual sort of entry with time stamps, user ID, and the device of the user. And then over here, we have a couple of support questions. It looks like we've had some people coming back and saying they're struggling to log in to the app. So to help us understand if this is really an issue and what's going on, I can add the account agent into the canvas, and I can use the selection option here to give the agent access to the support questions and the event stream. Hey, Kountai. I've had a few questions about, people struggling to log in to the app. I wonder if you could look at the support questions and see if it matches any data you can see in our event stream. Can you help me diagnose if this is a real issue, when it happened, or what we could do to fix it? Okay. Let's see what kind of I can come up with. Okay. Looks like the agent's finished. And you can see here it's given us a bit of a a a diagnosis of the problem. It says that there has been a real and significant authentication outage. Two error types have taken place, and it seems to be Android specific, particularly on this app version. And the root cause appears to be a server side authentication, service failure. This is all just coming from looking at behavior of the of the events, I assume. You can obviously look at this in more detail. We can break this down, see its actual analysis. You can see the the affected users by errors per day here. And everything we can see here, we can click on and see how the agents built it. We can see and follow its methodology, see the code it's written to sort of get to its conclusion so we can really trust what it's doing. And if we're happy with this, we now can take this report, share it with the rest of the team, bring them in here, have a discussion, and work out what we do.