For CSPs, the Chatbot Discussion Heats Up Around AI

Let's talk about chatbots for a moment. We've all used them before and, by now, most of us have accepted their limitations as the lesser of two evils (the greater evil being stuck in an endless IVR loop). Many users report at least some level of satisfaction with chatbots, particularly when handling common issues such as password resets, basic account information and other frequently asked questions. Yet the fact remains that traditional chatbots struggle with replicating a human interactive experience when the conversation goes off script. In those cases, it's often necessary to call in the live troops to resolve a customer's issue.

And that would be the end of our conversation about chatbots, except for the disruptive potential that AI brings to the discussion.

AI-powered chat agents have the potential to take automated customer service to another level. The ability to apply dynamic, human-like intelligence and real-time machine learning to customer experiences (CX) is what every chatbot has always dreamed of being. It is not, however, just a simple matter of bolting AI onto your existing chatbot app and waiting for better outcomes. There's a lot of business process integration and customization around industry-specific large language models (LLMs) that need to take place before a chatbot becomes an AI chat agent. Fortunately, that work is being done right now in an industry that can desperately use it: telecommunications.

Dialing Up AI for better Customer Experiences

Dell Technologies is honored to be part of an initiative led by SK Telecom to develop an AI chat agent application for communications service providers (CSPs). In addition to SK Telecom, which is developing the application, there is a critical middleware component being developed by Compax (on the business integration side) and Matrixx Software (on the billing integration side) to bring everything together. Dell's contributions center on three solution aspects: the cloud foundation for the application (via Dell Telecom Infrastructure Blocks), the underlying AI infrastructure (via Dell and Nvidia's AI Factory) and the multivendor solution integration and validation, which is being hosted in Dell's Open Telecommunications Ecosystem Lab (OTEL).

Part of the challenge of building an LLM-based AI chat agent for CSPs is the language itself. Telecommunications has its own set of industry-specific terms, from roaming fees to family plans. SK Telecom is currently building a customized LLM for CSPs that is capable of understanding the nuances of telco terminology. Beyond that, there is the challenge of integrating these conversations with back-office and billing systems that are often highly complex and, in some cases, have been in place for decades. And, finally, the entire solution needs to be deployed in a cloud-native architecture featuring AI-ready infrastructure. It's these individual components that brought SK Telecom, Dell, Compax and Matrixx together, with Dell's OTEL being the logical place to integrate and validate the entire solution.

Once completed, SK Telecom plans to roll the solution out to other CSPs around the world. For an industry that thrives on communication, AI chat agents are a potential game changer for telcos. Imagine, for example, being able to get real-time recommendations on plan changes and shop for new devices with an intelligent shopping assistant. Or get AI advice on choosing the best roaming plan right before you leave for your European vacation. These are the kinds of complex tasks that AI chat agents can perform and-here's the key-get better at over time by learning your preferences. Even human agents often struggle to connect on that kind of personal level.

We've come a long way from switchboard operators, and the best is yet to come. Want to talk about developing your AI app for telecommunications? Talk to Dell or check out more on the AI Chat Agent we are working on with SK Telecom. We're here to keep the AI conversation moving forward.