Telefónica has revealed its new artificial intelligence-based approach to customer service is a way of moving beyond traditional KPIs to better reflect user experience.
Earlier this week the operator revealed it was rolling out Huawei’s Smartcare Service Operations Centres (SOC) solution, initially across its mobile networks in Germany, Argentina and Chile. The solution will allow Telefónica to process large amounts of data in order to assess performance.
Speaking to European Communications, Juan Manuel Caro, Director of Operations & OSS at Telefónica, says the shift could change the network's entire approach to planning and maintenance.
“We are very used to monitoring the networks,” he says.
“We have lots of tools and KPIs to measure throughput, dropped calls and other things. When we detect a failure we try to sort it as soon as possible.
That is the traditional way of working and we are very good at it.”
This needs to change, Caro argues, because of the changing nature of telecoms services, which have changed considerably over the past decade.
“We are creating more complex services over the network, VoLTE, IP-TV, where we can have network KPIs all green [but] the customer can be having a bad experience.”
The only way to quantify the experience of Telefónica’s customers has until now been asking them through surveys. The telco uses this feedback to build its customer satisfaction index.
Upon rollout, Telefónica will use the Smart SOC solution to automatically find out what this score is in real time using anonymised, aggregated data from customers.
“When a base station is down we have alarms in that network. Now we can detect whether at the same time when a group of customers are having a bad experience.
“We will know that these customers are all served by the same base station, so we can tell from the customer experience that there is a problem with that base station.”
The new tool aims to address the problem of "silent churners" or customers who do not complain but experience poor service and eventually move to a different provider. It will also feed new information into Aura, its digital assistant.
The tool is capable of learning, Caro explains. To begin with, if there is a problem, the operator's staff will have to deal with it manually so it can understand how to do so. He says in time the tool will be able to automatically detect the issue and correct it without human intervention.
This AI-led approach will dovetail into Telefónica's network investment, by addressing what customers actually need, says Caro. Simulations will be run to assess how customer experience might be impacted if certain factors change.
For example, Caro says, WhatsApp users do not need high throughput. Video users only need enough throughput and bandwidth to the extent that the service does not degrade.
“In some places where I’m trying to get better KPIs I won’t need to put traffic there, because the capacity and the experience is already enough.”
As Caro explains, the telco had previously planned the network through purely technical KPIs.
“We cannot forget these,” he says, “but we have one more KPI: how these investments will impact customer experience.”
The hope is AI will lead it to smarter thinking.