Without context, AI guesses — creating unreliable outcomes. Ontologies aim to solve the problem.
During Mobile World Congress Barcelona 2026, Totogi CEO Danielle Rios sat down with RCR Wireless News’ Sean Kinney to get to the bottom of what’s stopping telcos from running AI agents, even when they have hundreds of them deployed.
“It’s all about governance,” she said. “They’re not sure they’re going to make the right decision, so they’re pumping the brakes.” The root problem, she explained, is context.
For example, different telco systems often define “subscriber” differently. Is it a device? Is it a line? A person or an enterprise? AI, while powerful at making decisions, struggles in these ambiguous scenarios. When it lacks context, it guesses, leading to the unreliable outcomes that operators worry about.
Rios said that until telcos model their entire operation — their business processes, the decisions they make, their rules — in a way AI agents can use across the enterprise, the problem will persist.
Totogi’s answer is an ontology — semantic framework that defines how data, concepts, and relationships connect within a domain. Unlike simple classification tools, they capture the relationships and context that allow systems to interpret and act on data, going beyond taxonomies or schemas that merely organize information.
The Totogi Ontology sits atop existing BSS, OSS, and network systems, unifying how your business is represented and enabling AI to understand, reason, and act across your entire environment. By normalizing entities, processes, and actions from siloed applications into a single semantic model, it creates a digital twin of your telco that allows AI to operate reliably and execute tasks with reduced hallucination risk. “It’s one representation of your telco, one truth,” said Rios. “When you create an AI application, it runs through the ontology and down to your legacy systems. The ontology won’t let it make a mistake.”
Real-world deployments
Totogi says several operators are already applying the approach in real-world deployments. Nearly 10 Tier-1 operators are currently using its ontology solution.
One customer, Zain Sudan, faced a network issue where certain cells would go dormant without triggering alarms. Before implementing the Totogi Ontology, identifying and resolving the issue typically took about 48 hours. “We brought that down to 30 minutes,” Rios said. “And now we can add predictive AI on top of it and stop it before it even happens.”
In contrast, Singapore-based StarHub is using the Totogi Ontology within its enterprise business unit to accelerate sales training and support live customer conversations. As sales representatives speak with customers, the system transcribes the conversation, suggests responses, and recommends relevant products. “It’s like having your CMO recommend the exact products you want to sell, bundling the products in the right way, with the highest margin,” Rios said.
The approach aims to preserve the human element of enterprise sales while giving representatives real-time guidance during conversations.
“Once telco starts to use that ontology and unlock use cases, we’ll really see a hockey stick of adoption because they can trust AI will make the right decision,” Rios said.
