- Calix teamed up with Google Cloud to launch agents tailored for broadband providers
- The goal is to make agents easier to deploy regardless of where telcos are at with AI
- Providers can use pre-configured agents or make their own
Agentic AI adoption is still in its early stages for most enterprises, telcos included. But Calix thinks it has just the tool to give broadband providers a head start.
Calix, with the help of Google Cloud, is upgrading its broadband platform with agentic AI capabilities, which will allow customers to deploy agents for tasks across operations, customer service and data insights.
The product, which will launch later this quarter, aims to make agentic AI more accessible to operators and also ensure that deployments don’t eat up operating expenses, according to Calix CEO Michael Weening.
“Our customers tell us all the time they don’t have enough capacity. There are a whole bunch of things that they like to do,” Weening told Fierce, like building networks and launching new services. Calix’s upgrade, which he said is the third generation of its broadband platform, is “a massive augmentation of their capacity.”
Layer cake
He explained Calix built a series of “layers” to help operators use pre-configured agents or make their own if they’d like.
First comes the data layer, which involves putting network and subscriber info in one place. The Calix platform then uses Retrieval-Augmented Generation (RAG) and Model Context Protocol (MCP) to make sure large language models (LLMs) are accurate and understand context – this is the knowledge layer.
The operator then gets to decide how much an agent can do for them, and that’s where the trust layer comes into play, said Weening. “Are they allowed to read, re-write only? Can they actually make changes?”
Establishing these kinds of parameters is pretty crucial as the telecom industry still has an unrefined risk mitigation process.
For Google, Calix’s agentic AI play is a chance to “really democratize” AI for the broadband industry, according to Angelo Libertucci, Google Cloud’s global head of telecom. One of the biggest challenges broadband providers face in deploying AI agents is not having enough talent or resources, he said, and they’re all in different stages of the process.
Some have data scientists to build their own AI models, while other telcos just want to get a better sense of the technology’s return on investment. Regardless of where they’re at, “they want and they need to do that quickly,” Libertucci added.
The Calix and Google Cloud collab comes as the latter aims to secure a bigger slice of the agentic AI market. Google this month introduced Gemini Enterprise, a business-friendly version of its AI assistant that notably works across non-Google applications. The cloud giant also plans to make its fiber network fully autonomous by end of 2025, so unsurprisingly it’s looking to help telcos do the same.
Weening said Calix currently has just over 1,200 customers, ranging from large Tier 1 telcos to rural cooperatives. But he thinks both big and small operators have the same questions about agentic AI deployment.
They’re asking, “who is providing me the best stack, who provides me trainable agents and what’s my fastest time to market at the highest ROI?”
Of course, Calix isn’t the only vendor doubling down on AI agents for telecom providers. ServiceNow earlier this year rolled out agent models specifically tailored for the industry. Additionally, Netcracker last month released a new iteration of its agentic AI platform that comes with more than 60 pre-built agents.