Murtaza Doctor

So this is what the power of agentic AI is.

And it's taking us a long way.

And there are obviously a variety of definitions here, which I would love to get into, but I kind of wanted to set the tone for us on what I believe what agentic AI is and how we are driving it at ThousandEyes.

Yeah, absolutely.

I mean, I think this loop is so powerful because not only

that you need to gather a lot of context very, very quickly, which an agent is very capable of doing.

You don't have to look for a needle in a haystack.

And then the reasoning part, which is the intelligence coming to the model,

And we don't want to stop there.

We actually want to act and do some things, which is, you know, in this case, it not only correlated everything together to figure out that, you know, what the actual issue was, it created what we call as an RCA, root cause analysis or a diagnostic node around it.

and sent it on Slack on the SRE channel for everyone to know.

And the important part which you mentioned is the MCP part, right?

And which is what we call as the model context protocol.

And the reason why this MCP really matters is because it is becoming the de facto standard for tool invocation.

Yeah.

So in this case, you could have access to your GitHub, your Atlassian, your anything under the covers, a database, a file system.

And it's basically like an agent to infrastructure API layer.

And I think basically you can think of it like OAuth or REST, but for AI agents accessing enterprise systems.

That's how I would like to call it.

Absolutely, right?