Unryo Rebuilds 5G Signaling into Sequence Diagrams and Marks Failures

Service-based interface analysis that locates each failed 5G procedure down to the network function and pod

Unryo has released agentic AI analysis of 5G core signaling, which reconstructs probe traces into per-procedure sequence diagrams and identifies where a procedure failed or slowed down.

Control plane troubleshooting in a 5G core is still largely manual. An engineer pulls a trace, works out which exchange failed or ran slow, and then checks the network function and the infrastructure beneath it in separate tools. The work is slow, it is repeated for every incident, and the result lives in one person’s notes.

Also Read: Top Data, AI and Cloud Conferences for CIOs and Data Leaders to Attend in 2026

Unryo’s AI agents do that work on request. An engineer asks in plain language for the signaling of a given subscriber identity (IMSI) over a chosen time window, at any time and with or without an open incident. The agents retrieve the matching traces and assemble them per procedure, including registration, PDU session establishment and handover, into a sequence diagram that shows every request and response between network functions in chronological order.

A failed step is marked with its response code and the network function instance that returned the error. The agents also flag missing messages, such as a lost packet, and use 5G protocol context to judge whether each response time is acceptable.

Also Read: IT Tech Pulse Exclusive Interview with Michael Jack Chief Revenue Officer and Co-Founder of Datadobi

When the agents find a problem, they use the platform’s topology to identify the infrastructure element serving that network function, down to the pod, and check whether that element is saturated, which would make it a probable cause.

Investigations start from the failing step and its probable cause, not from a raw trace, so engineers spend their time on the fault instead of on piecing the investigation together across tools. Every query the agents run is shown alongside the result, so the evidence behind each diagram can be checked rather than resting on one person’s notes.

Write to us [⁠wasim.a@demandmediaagency.com] to learn more about our exclusive editorial packages and programmes.