Unryo adds TM Forum Open APIs to agentic observability platform

Unryo says its network observability platform now uses TM Forum Open APIs to move alarms in, diagnosed incidents out and make its AI agents governable through open standards. The move is aimed at telecom operators that want less custom integration and more auditable automation across multi-vendor networks.

Why it matters: - Telecom operators are under pressure to connect more tools, more vendors and more AI without adding brittle one-off integrations. - Standard interfaces can let network data move between alarm systems, incident systems and AI agents with less translation work. - Governable AI is becoming a bigger operational and regulatory issue as operators let software make more decisions.

What happened: - Unryo implemented three TM Forum Open APIs for telecom operations. - TMF642 Alarm Management brings alarms into Unryo from compliant source systems. - TMF724 Incident Management sends diagnosed incidents out of Unryo to downstream systems. - TMF915 AI Management exposes Unryo’s AI agents as governable records. - The company says the approach replaces custom integrations with open standards.

The details: - TMF642 input lets any system that emits TMF642 alarms send them to Unryo. - Unryo normalizes and correlates those alarms with other sources before feeding its AI agents. - ServiceNow’s Alarm Management module and Ciena Blue Planet are cited as examples of TMF642-capable telecom assurance tools. - TMF724 output gives consuming systems root cause, correlated alarms, fault object and affected resources. - Unryo says downstream systems receive a finished incident, not a raw alert stream. - TMF915 lets an external system query what AI Unryo runs, the version and characteristics, and inventory those models alongside others. - Unryo says its correlation is deterministic and includes per-decision reasoning. - The company says the same TMF interfaces expose the same data Unryo uses internally, rather than a reduced copy. - That shared view is meant to keep the NOC, downstream systems and AI agents aligned on one fault picture. - Unryo’s platform is also built around a Topology Data Fabric that maps multi-layer dependencies across physical, logical and application infrastructure. - The platform is available on-premise, in the cloud and in hybrid deployments.

Between the lines: - The TM Forum stack gives Unryo a standards-based way to sit between many alarm sources and many incident consumers. - That structure fits the classic Manager-of-Managers model used in telecom operations. - The TMF915 piece is the more strategic shift because it treats AI itself as something operators can inspect and govern, not just use. - Unryo is also signaling that explainability and auditability matter as much as automation performance. - The reference to the EU AI Act suggests the company wants its AI posture to align with emerging oversight expectations, while saying compliance remains the operator’s job.

What's next: - Operators that already support TM Forum standards can connect to Unryo without building a separate connector for every tool. - OSS, ITSM systems and autonomous agents can consume diagnosed incidents through the same standard path. - The company is positioning the platform for broader telecom adoption as agentic AI moves deeper into network assurance workflows.

The bottom line: - Unryo is trying to make agentic observability interoperable, diagnosable and auditable at telecom scale.