Google and Nokia expand the use of Gemini AI agents in telecom network assurance

Google and Nokia are moving deeper into telecom operations with a new deployment of “AI agents” designed to help manage live networks. Yahoo Finance reported that Nokia will add six Gemini-powered agents to its Assurance Center, a software platform used by telecom providers to oversee network performance, detect issues, and support troubleshooting.

Gemini-powered agents are software programs that use Google’s Gemini models to perform specific, narrower tasks, rather than acting as one general chatbot. In a network-operations setting, that distinction matters because telecom teams typically need repeatable workflows, such as identifying anomalies, correlating events, and guiding next steps for remediation, rather than open-ended answers.

Under the reported plan, Nokia’s Assurance Center will incorporate six of these agents. Nokia did not outline, in the reported summary, the detailed scope of each agent, such as which assurance workflows they cover first or what level of automation they will perform versus what will remain human-led.

For Alphabet, the relevance is twofold: the announcement reinforces Google’s push to place Gemini technology into industry-specific environments, and it positions Google as a key upstream AI provider for mission-critical systems used by large telecommunications networks. For Nokia, the goal is to extend its assurance tooling with AI that can reduce the time between detection and resolution, and potentially lower the labor cost of network operations.

AI in telecom has been pursued for several years, but operators have increasingly focused on “agentic” systems that can take actions within defined boundaries. Assurance platforms are a natural fit for this approach because they already rely on structured indicates, monitoring data, and diagnostic workflows, which can be paired with AI to summarize what is happening and suggest operational responses.

Even with the reported partnership, it is still unclear how Nokia will validate the agents in production. Operational AI typically requires careful controls, such as confidence thresholds, auditing, and fail-safes, so that operators can trust outputs and avoid incorrect remediation. The reported account does not provide details on those governance mechanisms, nor does it specify whether the agents are initially deployed in “assist” mode, where they recommend actions, or in “execute” mode, where they can directly trigger automated steps.

Nokia also did not disclose in the report what customer segments will receive the update first, how quickly it will roll out, or whether the six Gemini-powered agents require new licensing or are bundled into existing Assurance Center contracts. Those commercialization details can materially affect adoption, since assurance platforms often sit at the center of operator workflows.