Artificial Intelligence for Telcos? Experts Discuss Applications, Impact and Possible Barriers
Thanks to ChatGPT, the hype around Artificial Intelligence (AI) has gone through the roof. So how should the telco sector think about the technology? Is it all about customer experience? Or should CSPs apply AI to networks and infrastructure first? Netcracker's Head of Strategy & Portfolio Marketing Sue White joined a TM Forum debate on the topic.
There is amazing potential, but there are also many challenges ahead for AI technology as it begins to seep into all areas of life and work. In this TM Forum panel session titled Counter Intelligence: Using AI to Improve the Customer Experience, four expert guests dived into the topic – specifically as it relates to the telco sector. Experts included:
- Dean Ramsay, Principal Analyst of TM Forum
- Sue White, Head of Strategy & Portfolio Marketing at Netcracker
- Matt Sanchez, Global Chief Data & AI Officer, Tecnotree
- Teresa Cottam, Chief Analyst, Omnisperience and TM Forum
Cottam began with an overview based on a forthcoming research paper on AI in Customer Experience and Customer Service, which is to be published by TM Forum. She reflected that AI is currently at the peak of the hype curve, accruing an estimated $154 billion in spend this year (growing to $300 billion by 2026). Cottam also referenced a projection that the telco sector alone will spend $40 billion on AI by 2031.
Generally, the panel agreed that CSPs are being relatively cautious in their approach to AI and are behind other industries. . Although the technology is being used across all areas of the telco business today, most are not yet truly data/AI driven whereby data is an asset to be monetized with a strategy driven from the top down. This is understandable given the embryonic nature of the technology and the high risk of project failure.
Still, plenty of work is underway. Sue White noted that in addition to customer experience, there is a lot of focus on operational efficiency. She said: "We’re seeing projects across the entire spectrum, and the ones that make the most sense are those that are designed to bring down OPEX cost– things like anomaly detection and predictive maintenance that will reduce churn and calls to the call center."
ChatGPT and other large language models (LLM) are likely to be the next major breakthrough in AI-based CX. While chatbots have become the norm in customer care, they are also a source of frustration for many customers. LLMs, trained on telco data and with access to customer history and context, will revolutionize the chatbot experience by taking on more complex problems and engaging with customers in a more ‘human’ way.
While LLM integration will evidently reduce the need for customer care agents, it will also provide additional support for agents to solve problems that need real human interaction or help speed up back end processes. . As an example, Cottam noted call transcripts, which take a lot of time to write and can be tedious for call handlers. An AI could step in and make intelligent notes, saving a huge amount of effort.
Transforming into a data-driven business has become a core strategy for CSPs, but there are considerable technical challenges to overcome for any large enterprise that wants to embed AI across its organization. Most obviously, finding the training data, reading it and combining it into something useful.
White said: "You have to get the data right. When you do, it can be really transformative. But it's difficult. Data is stored in many places. And there are complex IT environments that have not yet been modernized. Being able to extract it is one thing, but transforming it to a usable format is another.
"For example, TM Forum's SID format is great for BSS/OSS, but it's not useful for analytics systems. So at Netcracker we've been working hard to make sure data from many sources is available in usable form – and when we say usable we mean for different parts of the business that don't have IT expertise."
Netcracker’s Data Analytics Platform breaks down those existing data silos. It combines data management, advanced analytics and smart use cases – such as reducing the churn rate or preventing fraud – to help telcos harness and process the right data. This makes it easier for all business departments to create useful insights with self-service analytics.
As White explains, “they can now get the information they need on their own dashboards” – whether that’s in marketing, product management or service operations.
You can watch the entire TM Forum session here.