Richard Thurston, TelecomTV (00:02):
We're here with Osama Ahmad and Emad Damra, and we're going to talk about Ericsson's end-to-end holistic services approach and the Ericsson Intelligent IT Suite. So Osama, tell us what you've got for us today.

Osama Ahmad, Ericsson (00:16):
Yeah, so the Ericsson Intelligent IT Suite is the enabler to move from automation to autonomy, and in the age of continuous transformation that CSPs are going through to modernize, monetize, and modularize their networks, they face a lot of familiar challenges, challenges from the multi siloed approach, challenges about waterfall approach in the transformation, making rigid systems come together and making it costly to transform, disrupting customer experience. And in the end, they're always falling behind the customer demands and the market demands. So what Ericsson has done to make these transformations successful is we've brought together all this complex ecosystem under a single pane of glass with a telecom knowledge, transforming the infrastructure and CI/CD as a single source. We'll go into a bit of the services here and show you what this suite consists of. So it consists of our holistic services approach, as you mentioned from the Advise, Build, Operate, and Ensure. And the Advise starts with exactly mapping out what we need to do and then we Build it using Agentic AI, Operate it through intent-based operations and Ensure the services for predictive analysis. This is built on intelligent core where we have enablers through our assets, processes and technology and techniques and capabilities. I'll invite Emad to walk you through the demo.

Emad Damra, Ericsson (02:01):
Thanks Osama. So I want to show how we are accelerating the journey of autonomous operations to our customer and this is what we build on top of Intelligent IT Suite using our managed IT operations tools. So in the demo part, we start with where TM Forum did their autonomous network blueprint, identifying the intent across business, service and resource layer. In this demo we show more specifically about how to register the intent for specific case and how to initiate those multi AI agents and then how we'll perform impact analysis for some cases. Of course, if we start by the intent part, there could be couple of different business and service intent and those could vary between different issues and needs for the customer. We'll focus about Peter. Peter here is launching a 5G offer and campaign for the consumer enterprise and this campaign is expected to have a higher rate of order failout in the system.

(03:18):
So Peter now will be chatting with the generative AI that we have as a chatbot to start creating and registering this intent. So by chatting with the AI now we are enabling our customer to register those new intent for order fallout handling. The AI of course will ask about the detailed of scope of how to handle this intent in the observability part on the operation. Of course, Peter is sharing the information about the start date and what to handle exactly. Then more detail about the scope of how to handle such order failures for specific cases. So now the agent is of course confirming the scope of that intent and based on that it will schedule it. So what will happen, the different multi-agents will be created to handle this campaign and we'll see that in the next screen where we will have our Observe Agent, Prediction Agent and Decide Agent.

(04:23):
Those work together to, first, Observe Agent is monitoring the orders happening in different system Prediction is analyzing that in real time, predict if there is any issue could happen. Then Decide and Actuation agent, create the proposal and resolve it based on the root cause analysis. So let's take example. The campaign started, it's 16 December orders are coming and now at 11:00 AM we can see that the infrastructure utilization is start to increasing. So the prediction agent is now expecting that this will really have a surge on the order management system in the next 30 minutes. So it's advising the site agent to recommend to escalate that to L2 and automatically for raising another creating ticket for L2 to address it. If we take another example, and this is where we detect the spikes in order. So the site agent, based on the knowledge that they know they suggest to try and solve it as well based on the actuation agent. So those agents work together as a conclusion based on the knowledge pane and management that we have, they learn as well from human interaction and eventually supporting our aim to deliver intent driven operation and AIOps predictive operation.

Richard Thurston, TelecomTV (05:49):
That's really interesting. Thank you both.

Emad Damra, Ericsson (05:51):
Thank you.

Osama Ahmad, Ericsson (05:52):
Thank you.