AI and Chatbots within telecoms: Q&A with industry analyst Teresa Cottam
The TM Forum recently published a report on utilizing AI in telecoms customer service. Subtonomy’s Tina Rosen caught up with the author, industry analyst Teresa Cottam, to find out how CSPs are using AI and what the next big trends are going to be.
#1. How big a deal is AI in telecoms customer service and how advanced are CSPs in implementing it?
Hi Tina. AI is a very big deal for telecoms CSPs who are looking to implement it across their organization but particularly in customer support. Currently the most familiar manifestation of AI/ML is the chatbot. Lots of CSPs have rolled these out, with mixed results. The first generation of chatbots, in particular, were often very frustrating for customers as they weren’t very smart! But the technology is evolving rapidly and CSPs are excited about the next generation of chatbots they’re bringing to market. Beyond chatbots though, AI is also being applied within key customer service processes to deliver smarter self-care and greater efficiency in the call center. But despite all the excitement, CSPs are not on the cutting edge - they’re trailing other industries.
#2. From your research, what challenges do CSPs face when implementing chatbots?
Three common pitfalls spring to mind. The first is that chatbots are sometimes presented as though they’re a panacea. While they’re a useful component of the service mix, they have limitations. They’re not going to be able to handle every inquiry and not all customers are going to want to interact with them. The second challenge is to do with data.
Telecoms has huge amounts of data, although we’re not always very good at utilizing it efficiently. But data is the fuel that chatbots and AI-infused operations need to work efficiently, so we’re not going to be successful with these initiatives unless we have the right data foundation.
The third main challenge is how we’ve gone about implementing AI. It can’t be implemented as a silo, delivered by technologists who have no real knowledge of the business or of customer needs. If CSPs do that they risk implementing a capable but dumb tool that no-one wants to use because it doesn’t solve real problems and just gets in the way.
#3. How do CSPs optimize their chances of success?
They have to get the foundations right by bringing all the data together so it can fuel AI-empowered omnichannel service. They have to involve business users who understand the context of the business problem by making it easy for business users to utilize AI to solve real customer service challenges. This helps maximise the ROI which is very important, because you need to show quick wins. I think my final tip is to really think about the KPIs used to measure success. A Hubspot study found that 40% of customer service leaders still see customer service as an expense rather than a growth driver and this is reflected in the fact that the purpose of the first generation of chatbots and AIOps was purely operationally-centric – to save money and deflect calls from the call centre. The result was a lot of customer frustration.
Choosing the right combination of KPIs and including customer-centric KPIs is really important. Think about how AI helps you boost customer lifetime value, reduce churn or increase customer satisfaction, not just how it saves service costs.
#4. What’s coming next in the chatbot world?
Wow that’s a big question as so much is happening. I think we’re going to see a lot of activity with AI/ML being infused into more processes and technologies and used in combination to create more capable solutions. When conversational AI is combined with sentiment analysis and predictive analytics, for examples, chatbots move beyond searching for keywords and providing standard answers to processing natural language and customer intent. I think the goals or purposes of these project will also change – moving beyond cost control and call center deflection to a specific goal of enhancing the service experience. And I think we will be using AI to tie things together into a more coherent experience. Chatbots will work from wider and deeper datasets. Data will help chatbots personalize the experience and sentiment analysis will help them adapt to customer moods and emotions in real time. We’ll start to see an emphasis on convenience and effectiveness from the customer’s perspective.
The AI, for example, will assess if the customer is getting a good service experience and will route calls effortlessly to human service agents if needed. The human agent will then be able to seamlessly pick up the service call because they’re working from the same dataset and will have all the history and key issues in front of them in their applications.
I also think video chat will become more common because it delivers a level of authenticity and engagement that voice doesn’t. It’s already being used in other verticals – Salesforce, for example, said that video chat usage has increased 47% since 2020 - and you can see that it would have powerful applications in B2B telecoms in particular. AI will help here by analyzing video in real time and logging insights from calls. It’s only a relatively small step from video chat to VR-based customer service when chatbots evolve into avatars. We’re already experimenting with that. Capita, for example, rolled out a 3D digital avatar in 2022 which shows emotion and empathy and reacts to customer sentiment.
#5. Gen Z and chatbots. Are they made for each other?
Absolutely. For a long time, we talked about Gen Z as the next big thing. Well, that next big thing is here. They comprise 1.8 billion people, and will represent 75% of the global workforce by 2025. They also have a combined spending power of $2.5 trillion per year. So, Gen Z are the majority of our workforce and the majority of our customer base. When we think about building customer support, these are the customers we’re building for. LivePerson did a study that found 80% of those aged 18-24 feel positive about engaging with a chatbot and 60% of this age group would prefer to engage with a chatbot than a human. So, what do they want? They want fast answers. They don’t like waiting, repeating or length. Their emphasis is on execution. And there’s still a big gap between expectation and execution. That same LivePerson study found that 85% of companies think AI is important to their customer engagement strategies, but only 43% of customers think it’s currently easy to talk to chatbots. That’s a significant gap. CSPs need to focus on closing that gap and ensuring the next generation of chatbots are efficiently delivering both the answers and actions Gen Z need.
Discover more - The ultimate solution for technical customer support
Subtonomy offers the most extensive solution for technical customer support tailored for the telecom industry. We help Communication Service Providers (CSPs) to meet today's challenges and ever increasing demands and deliver an outstanding customer service and support with a unified customer-centric view.