logo_header
  • Topics
  • Research & Analysis
  • Features & Opinion
  • Webinars & Podcasts
  • Videos
  • Event videos
topic

Sponsored

How autonomous networks are elevating experience and unlocking value

In this interview with TM Forum, Ericsson’s Mats Karlsson discusses how to make impactful changes towards autonomous networks and autonomous operations.

Mats KarlssonMats Karlsson
23 Jun 2026
How autonomous networks are elevating experience and unlocking value

Sponsored by:

Ericsson

How autonomous networks are elevating experience and unlocking value

Mats Karlsson, Head of Solution Area Business and Operations Support Systems at Ericsson, which is a Platinum Partner of DTW Ignite 2026, outlines how agentic AI, autonomous operations and a “knowledge plane” can help CSPs simplify complexity, accelerate monetization and scale decision-making safely.

TM Forum: What can we expect to see from Ericsson at DTW Ignite this year? Which sessions are you participating in, for example, and what will you be showcasing on your booth?

MK: We’ll be in the main hall, the Ericsson booth is #303. Come by and see us there!

We’ve several OSS/BSS demos along with network automation and AI and managed network services. Progressing towards autonomous networks is our big focus. For OSS/BSS, we’re demonstrating how to: “Sell. Deliver. Get paid”, “Boost OSS/BSS with agentic AI” and how to embark on our “Business Value Pathways” to make measurable progress towards autonomous networks and in high-value scenarios.

From network automation and AI we’re showcasing four case studies in RAN automation, the Ericsson Intelligent Automation Platform (now for Core network too) enabling cApps and many use cases for automation in the rApps portfolio.

We’ll also be showing how Ericsson Operations Engine supports AI-driven closed loop operations in high value scenarios.

As well as the Ericsson booth, we’ll be at our partner AWS’ booth showcasing OSS/BSS on AWS for “Intelligent operations for charging and billing”.

And you’ll see Ericsson on several of the conference stages:

And I’ll be on stage twice:

  1. As part of the Autonomous Networks Leadership Forum with Indosat Ooredeoo Hutchison’s Chief Digital Officer Sukhdeep Gill in Indosat Ooredoo Hutchison’s outcome-based transformation on the road to autonomous networks
  2. In panel discussion on adopting agentic AI at scale: AI Native ODA - the path to open digital autonomy

TM Forum: You are also heavily involved in the Catalysts again this year. Can you briefly explain which projects you have joined, and why?

MK: The Catalysts offer us a unique opportunity to collaborate with customers (and often our competitors!) to combine our experience and wisdom in the pursuit of progress. You can see us in two of the Moonshot Catalysts: Conflict management in intent-based networks - Phase II showing how conflict between autonomous entities can be avoided, detected and resolved; and Business-aware GNN-healing networks empowering self-driving networks with graph AI and digital twins for predictive root cause analysis and Level 4 autonomy.

We’re also working on the Catalyst project Telco AEGIS: Autonomous ecosystem for generative intelligence and security, transforming CSPs into resilient digital service providers that are better equipped to protect customers, meet regulatory expectations, and sustain growth.

In addition, we will be at the Innovation Hub with an agentic AI platform with autonomous negotiation, fulfilment, and continuous assurance capability enabling L4 network service contracts — from offer to operation.

TM Forum: You focus on Ericsson’s OSS/BSS activities. What are the biggest challenges your customers are facing here as they move towards autonomous networks and operations?

MK: While the hype surrounding artificial intelligence (AI) continues to grow, siloed data remains a major roadblock, preventing AI from reaching full production at scale.

Furthermore, progress toward autonomous networks is fragmented. While maturity models exist, many CSPs still don’t feel the difference in their operational reality.

Progress toward higher autonomy will greatly enhance operational efficiency for existing services and is essential for enabling differentiated connectivity and new business opportunities at scale.

Five fundamental barriers must be overcome:

  1. The shift to autonomy: Existing OSS/BSS environments must evolve to support autonomous operations and the demands of mission-critical networks.
  2. Time to monetization: We want to make it easier to innovate and capture value.
  3. Operational complexity: We need to unify and simplify interfaces and improve issue resolution accuracy.
  4. High costs and scaling challenges: Manual processes drive up time and cost to resolution, making scale difficult without significant resource expansion.
  5. Data bottlenecks: We need the right data in the right form at speed to support intelligent decision-making.

TM Forum: How is Ericsson addressing these challenges with your customers, and how will AI and agentic AI help?

MK: The shift toward autonomous networks requires a fundamental shift in strategy. Your OSS/BSS transformation shouldn’t just be about keeping up; it should be a continuous pursuit of better business outcomes.

To really transform, focus on outcomes that matter rather than technical achievements. We’re introducing four initial OSS/BSS focused Business Value Pathways towards autonomous networks goals each of which carries commitment to transform the customers autonomous network maturity of key high value scenario capabilities:

  1. Faster data transformation: Underpin autonomous operations with the right data in the right form in real-time, so you move from issue prevention to new revenue without the overhead.
  2. Zero-touch product launch: Automate product launch.
  3. Agentic AI service experience: Agentic AI-enabled shift from reactive troubleshooting to assisted and accelerated issue resolution.
  4. Agentic AI intelligent IT operations: Give engineers a unified operations interface to resolve faster and protect experience.

Throughout our portfolio we’re continuously evolving, getting CSPs ready for autonomous operations, to:

  • Accelerate monetization with intelligent charging and billing for multi-sided models and experience-based pricing.
  • Automate service orchestration to turn intent into action. Add real-time inventory so the network can self-heal, self-configure, and self-optimize.
  • Plan for an agent fabric that makes automation and AI trustworthy and safe for closed loops at scale. Here is a link to a recent blog on this topic: Accelerate automation in OSS/BSS with agent fabric

TM Forum: You have talked about the “knowledge plane” concept as the data foundation for autonomous networks. Can you explain what this is?

MK: In collaboration with Telstra, we recently published an industry perspective: From automation to autonomy

We’re collaborating to define the path forward. The combination of intent and knowledge is the foundation for decision-safe network operation. Intelligent automation is an important transitional capability, but not the end state.

The systemic risk associated with making changes leads to what we define as the fragility trap. To prevent failure, change is minimized, resulting in operational inertia, rising costs, and reduced efficiency. Breaking this cycle requires decoupling change from instability.

The fundamental shift lies in intent-based network autonomy. Define desired outcomes and the system autonomously plans, designs, validates, and executes changes.

We need to enable predictable, “safe-by default” operations at scale. Engineers transition from operating within the loop of execution to governing the system itself (with agentic AI) defining policies, setting intent, and overseeing outcomes.

While intent provides direction, the knowledge plane provides the context required for safe decision-making. It encapsulates the information necessary for reasoning. It’s dynamic and machine interpretable. It enables systems to validate decisions at runtime, ensuring that actions remain feasible and compliant with defined guardrails. At a minimum, the knowledge plane must represent entities, relationships, constraints, and governance structures.

Over time, it should evolve to incorporate learned insights and inferred knowledge, further enhancing system capability. Achieving Level 4+ autonomy requires more than increased automation. It demands a structured knowledge plane capable of supporting safe, consistent, and explainable decision making.

By managing the lifecycle of the autonomous system rather than individual changes, CSPs can simultaneously achieve scale, reliability, and efficiency — a trilemma that has eluded us historically.