While at Splunk .conf25, I had the opportunity to chat with Jason Allen and Tim Colpo from Accenture about their Continuum Control Plane solution (powered by Splunk) - a.k.a. CCP. CCP introduces an agentic approach by creating a data foundation that connects siloed, cross-domain data and provides a semantic layer on top. It also lays the foundation for AI agents that help users gain insights, inform decisions, and execute agentic actions. All of this is made possible by Splunk’s ability to aggre

Transcript

Auto-generated captions, lightly cleaned. Speakers are not separately labeled.

Hey everybody, we’re out here at docom 25 and I’ve been talking to Jason and Tim from Accenture partner of the year and we’ve been working a lot together on Agentic AI and CCP. Can you give the audience a little bit more information about what CCP is and what it does? >> Certainly, we we can certainly do that. So CCP is Accentur’s continuum control plane which is really right now transforming itself into an Aenteic platform that that connects lots of data sources and lots of data. Specifically, we really use Splunk to connect into CCP to drive some really cool use cases around root cause analysis, some financials, some some some different different ways to look at your cloud spin and and and really cool things. The really good part about it is is as Splunk came out with their MCP, they’re doing the standard protocol. We’ve started to plug that in and and it’s really working out well on on some of our newer use cases as as we drive in and use the different tools there. Now, what it is is like I said, an agentic platform. What it does is it looks at the different personas, who’s actually doing the work, how how they do the work, how they look at doing their work, and we start to create the agents around that >> drive the efficiencies in their work. Now what what Tim here is is really when people on the ground building agents and driving what we’re doing at Accenture. So Tim has created a a nice little root cause analysis. >> Yeah, we’ve got um we basically took um a manufacturing demo and basically pushed it into ITSI and we just built a a kind of your stereotypical uh simulated data, but what we’ve done with that is we’ve uh incorporated that with some of the or out of the box integration the service. now, but then now that we’ve plugged in um the Splunk MCP server and built with CCP and CCP’s um also built on the AI refinery, which is kind of the SDK that Esentia has to kind of layer the agents. So using that, we’ve come up with an agentic framework and different personas to do inquiries onto you know what is my service, what is my service score, um even plugging in some um custom service now capabilities to update tickets as well. So we’ve got a manufacturing demo. We’re also using rag and the MLTK and some other type of persona examples as well. We have over in our desk, sorry, our booth. We also have the uh like some enterprise security hits. But it’s a good demo and I’ve also had a lot of opportunities here to understand some interesting questions we’ve had and I’ve had to go over to the AI desk and say can do when is this coming out? So, it’s been actually very good conference to just have the ability to kind of have everybody um at their at your fingertips to kind of answer a lot of interesting questions. >> Yeah. >> And we were talking the other day about the MCP server with Splunk and Service Now, and you did it pretty quickly if I recall. No, it was it was plugged in pretty quickly, you know, and you know, you can use the ref refinery and get your MCP servers up, but it’s kind of having the strategy to work out how you want to use it because anybody can just bring up the the chat GPT or whatever else want to do and kind of do I mean the example I’ve been saying in the booth is I can say what is the status of my airline and you get like what the hell are you asking about? So, you’re kind of like starting to steer uh the MCP to well, what is my status of my service? and you kind of point it towards it. So you got to try to train the model and a lot of the work we’ve had to do because it is new is trying to make sure how we can get the model trained to kind of make it more efficient and and faster for the response. >> Yeah. >> So that’s a a lot of great information, a lot of work that you’re putting in. Now where can I go get some more information if I wanted that? >> So you can look at the link below or contact me at Accenture jc.mallccenture.com. [Music]