TPG wants customers to trigger AI agents instead of calling for help

TPG Telecom wants customers to put AI agents to work diagnosing and fixing service problems themselves, potentially skipping the call centre altogether.

That is the ambition outlined by general manager of technology operations Gaston De Arriba at Cisco Live in Melbourne, according to an iTnews report published on 12 November 2026. The report describes a capability under development, not a customer launch, and gives no rollout date.

The proposed workflow would let customers trigger agents for their own services, see service performance on screen and get guidance on improvements they could make. It could also create a trouble ticket without requiring a phone call.

“Our appetite is to give the power to the customer to trigger the agents from their own services, and then the agent to take care of the service performance for these customers,” De Arriba said, as quoted by iTnews.

For now, the agents are working behind the scenes. According to the report, they gather and correlate network performance data, key performance indicators, incidents, changes and complaints to help contact centre staff understand a customer's problem in real time. They also recommend what staff should tell customers and possible network fixes.

De Arriba claimed churn had fallen by around 40 per cent in the cohorts where TPG was using agentic AI. That is a company-reported result relayed by iTnews, not independent validation. The supplied report does not detail cohort sizes or the measurement period.

He also said proactive action was reducing incoming calls, sometimes addressing issues before customers noticed them.

The groundwork predates the customer-facing ambition. In a primary-source press release dated 4 March 2026, Cisco said TPG had deployed Splunk Enterprise, IT Service Intelligence, SOAR and Enterprise Security to support an AI-driven Service Operations Centre.

Cisco described the platform as bringing network operations and security together, correlating network data with IT incidents and using machine learning to predict and resolve service problems. Those are vendor statements about the deployment and its intended benefits.

Separately, iTWire reported on 15 April 2026 that root cause analysis previously taking 15 minutes to several hours was being completed in a few minutes in most cases, according to De Arriba. That measure concerns diagnosis, not necessarily the time needed to fix a fault.

Opening the agents to customers brings another set of risks. De Arriba told iTnews that TPG was moving cautiously, with defined boundaries around agent behaviour. Customer experience and security were priorities, alongside the “tokenomics” of running more AI workflows.

The goal is less time on the phone. The challenge is getting there without creating fresh service problems or an escalating AI bill.