At this year’s Mobile World Congress Barcelona, the telecom AI narrative shifted from copilots to collaboration—between machines.

Assurance specialist Mycom announced a strategic partnership with US-based cloud-native networking provider Mavenir to jointly develop Agentic AI use cases for 4G and 5G networks. The goal: move communications service providers (CSPs) beyond dashboard-driven monitoring toward semi- and fully autonomous network operations.

If the industry’s automation rhetoric is to be believed, that’s the promised land.

From Monitoring to Multi-Agent Autonomy

Traditional OSS (Operations Support Systems) platforms have long provided visibility—alarms, performance metrics, fault tickets. What they haven’t reliably delivered is autonomy.

The Mycom–Mavenir partnership is built around Agent-to-Agent (A2A) integration, connecting Mycom’s GenAie NOC Copilot with Mavenir’s Core Domain Intent Agent and domain-specific network function (NF) AI agents inside the mobile core.

Rather than a single AI assistant advising human operators, this model enables multiple specialized agents to collaborate directly—sharing context, diagnosing issues, and triggering remediation workflows across the network stack.

In practice, that could mean:

• Automated detection of network degradation

• Cross-domain root cause analysis

• Closed-loop remediation without manual ticket escalation

The architecture leverages Mycom’s PrOptima (performance management), NetExpert (fault management), and ProAssure (service quality management) platforms alongside Mavenir’s cloud-native mobile core intelligence.

The implication is clear: instead of humans stitching together insights across tools, AI agents do the stitching—and increasingly, the fixing.

Why Agent-to-Agent Matters for 5G

5G networks are inherently more complex than their predecessors. Virtualized cores, distributed architectures, and dynamic slicing introduce operational variables that strain legacy assurance models.

Manual workflows simply don’t scale.

By enabling secure agent-to-agent communication between OSS-level intelligence and domain-native core agents, Mycom and Mavenir aim to create a blueprint for structured, multi-agent collaboration inside live production networks.

According to Mycom Co-founder and CTO Mounir Ladki, the partnership is focused on operationalizing agentic AI “at scale,” not just experimenting with proofs of concept.

The distinction is critical. Many CSPs have tested AI pilots in isolated domains, only to struggle with integration across broader operational ecosystems.

A structured multi-agent framework, if executed cleanly, could reduce the friction between performance management, fault detection, and service quality enforcement.

Accelerating Toward TM Forum Levels 4 and 5

The collaboration also ties directly into the industry’s push toward higher levels of autonomy as defined by the TM Forum Autonomous Networks framework.

Level 4 and Level 5 autonomy envision networks that self-diagnose and self-optimize in real time, with minimal human intervention. Few operators have reached those stages at scale.

Mavenir’s EVP and CTO Bejoy Pankajakshan described the joint initiative as evolving OSS from a monitoring layer into a true autonomy platform—where assurance and optimization occur automatically and continuously.

That framing reflects a broader shift in telecom strategy. As 5G monetization pressures mount and operating margins tighten, CSPs are looking to automation not just for performance gains, but for cost containment.

Autonomous remediation reduces mean time to repair (MTTR), limits service-impacting incidents, and can lower operational expenditure. The business case is as much financial as technical.

Complementary Strengths

The partnership appears strategically aligned.

Mavenir brings deep domain knowledge in the mobile core, along with AI-driven networking solutions embedded directly into cloud-native architectures. Its agents operate close to the network functions themselves, enabling granular insight and closed-loop control.

Mycom, by contrast, sits higher in the OSS stack, offering end-to-end visibility across performance, faults, and service quality.

The agent-to-agent integration effectively connects domain-level intelligence with cross-network orchestration. That layered approach could help CSPs extract more value from existing OSS investments rather than replacing them outright.

In a market where rip-and-replace transformations are both risky and expensive, augmentation through AI may be more palatable.

Industry Context: From Assisted to Autonomous

The telecom sector has spent years pursuing automation through scripts, RPA, and rule-based systems. But rule engines break under the variability of modern, software-defined networks.

Agentic AI introduces a more adaptive model—systems capable of reasoning across context, collaborating with other agents, and taking action based on evolving conditions.

Still, real-world deployment raises questions:

• How are agent decisions audited for compliance and reliability?

• What guardrails prevent cascading automated errors?

• How seamlessly do agents integrate with legacy OSS environments?

These are non-trivial challenges, particularly in live 4G/5G networks supporting millions of subscribers.

If Mycom and Mavenir can demonstrate stable, secure multi-agent operations in production environments, they may provide a credible roadmap for CSPs aiming to reach higher autonomy levels without destabilizing operations.

The Bigger Picture

Telecom operators are under pressure to do more with flat or declining revenue growth. Network complexity is increasing, while tolerance for outages is decreasing.

Agentic AI offers a compelling narrative: networks that detect, diagnose, and fix themselves in near real time.

But autonomy in telecom isn’t a single leap—it’s a series of coordinated integrations across domains.

By formalizing agent-to-agent collaboration between assurance platforms and core network intelligence, Mycom and Mavenir are betting that the path to autonomous operations lies not in one super-agent, but in structured cooperation between many.

If successful, the partnership could mark a meaningful step toward Level 4 and 5 networks—where assurance isn’t just monitored, but executed automatically.

Get in touch with our MarTech Experts.