Ericsson injected agentic AI into its enterprise-focused NetCloud platform in a move targeted at helping taxed IT teams better manage highly variable 5G network connections across enterprise wireless WAN and private 5G deployments, a technology focus that the vendor said separates the offering from similar updates recently posted by market rivals Hewlett Packard Enterprise (HPE)-Juniper Networks and Cisco.

The agentic AI integration is through the NetCloud Assistant (ANA), which will gain several “agents” over the coming months. The first agent to launch will be a troubleshooting agent that will include automated workflows trained on issues identified by Ericsson support teams, partners, and customers.

Pankaj Malhotra, head of wireless WAN and security at Ericsson Enterprise Wireless Solutions, explained this “phase one” agent launch, which is set for later this year, builds on basic machine learning to provide an IT team with actionable insight into network performance.

“It's very focused on day-two operations, which is troubleshooting a network, and you could call it AIOps for 5G-based networks,” Malhotra said. “The real focus is to be able to assist that administrator who has built up that network using Ericsson-Cradlepoint devices and to be able to go back and figure out what is going on inside the network if there's an anomaly.”

Malhotra noted that this 5G angle is important due to the more variable nature of cellular-based network access.

“Wireless is a shared medium, and so because it's a shared medium, as more traffic comes in, you can expect your network performance to vary. It's not predictable,” Malhotra said. “So what it would be able to do is to go back and predict, based upon the information that is already gathered, based upon the traffic flows that are going through the system, about what you can expect out of that link. And then how can you go back and smoothen it, smoothing it in advance, to be able to go back and apply quality of service, but that happens when you actually know what the network link is going to look like.”

Malhotra said Ericsson expects this feature to reduce downtime and customer support cases by more than 20%.

Future AI agent updates will focus on configuration, deployment, and policy agents. These will bleed into some of the other NetCloud offerings, including secure access service edge (SASE), and also into Ericsson’s Private 5G platform.

Does Ericsson have a 5G advantage?

Malhotra noted that Ericsson’s wireless and 5G focus is what sets this latest agentic AI push apart from rival offerings. This includes Hewlett Packard Enterprise’s (HPE) updates to its Juniper Networks-based Mist and Marvis platforms and Cisco’s latest platform refresh.

“We focus really on that aspect because that's where we do believe that people shy away from those 5G deployments at times, because it is complex,” Malhotra said. “To be able to go back and bring that knowledge and that capability into an existing workforce, which is used to building up networks using wired links. By focusing on that, we believe that we are solving that problem first, where the collective knowledge that we have gathered over years for how to troubleshoot, how to configure, how to deploy wireless WAN-based networks is now condensed in these AI agents, which are sitting up there and trying to go back and help make the life of a system administrator easier.”

Ericsson is also tightening that 5G focus by bringing its private network portfolio under the NetCloud Manager.

“Till now, that private network portfolio was running under a separate network management system,” Malhotra said. “With this, we bring it under the same fold, which means that all of the capabilities that we are developing, including ANA and our agent architecture, now applies to our private network deployments as well.”

Ericsson’s escalating enterprise integration

The new offering also bolsters Ericsson’s growing enterprise focus. That effort accelerated with the $1.1 billion Cradlepoint acquisition that closed in late 2020, and hit an integration inflection point late last year.

Ericsson earlier this year bolstered its NetCloud SASE platform with a clientless zero-trust network access (ZTNA) option that allows customers to connect third-party and bring-your-own-device (BYOD) users. That move followed recent hardware- and software-related updates to the equipment providers Cradlepoint offerings.

Ericsson’s latest update comes as analyst predict surging revenues tied to the growing use of AI in branch architectures. Dell’Oro Group recently noted enterprise AI prioritization helped boost spending across SASE, SD-WAN, and SSE during the second quarter.

“With SASE revenue climbing 22% [year-over-year during the second quarter], enterprises are clearly prioritizing AI-ready branch strategies,” Mauricio Sanchez, director of enterprise security and networking at Dell’Oro Group, wrote. “SD-WAN’s 27% rise and SSE’s 19% gain show both the networking and security SASE engines driving this re-acceleration. Vendors delivering both SD-WAN and SSE capabilities now generate 84% of SASE revenue, reflecting the market’s confidence in integrated approaches.”