Authored by Justin Reilly, CEO of Wavelo
Think about how a customer most often interacts with their telecom provider. Your first thought might be when they buy a phone or set up a new plan. But that only happens once a year or every two years. The fact is, the most frequent interaction consumers have with their telecoms is about billing. In my experience, about half of all calls or contact with a telecom service provider concern billing. Even more than that are issues related to billing. Think about all of the different ways consumer behavior is telegraphed through billing related activities:
-
Increased data usage might signal a growing business.
-
Plan expansions could mean new family members coming online.
-
Global roaming might signal market expansion opportunities.
-
A sudden rush of refund requests could point to service interruptions (or new competitors).
-
An increase in billing cancelations could indicate a looming churn problem.
The list goes on…
All of these issues are related to and, most importantly, attributed to one system of record–billing. But today, all of that information is being spread across an increasingly jumbled mess of siloed operations like account management, marketing, customer service, and tech support, to name a few. So how do telecoms leverage all of this information?
The popular answer these days is to throw AI at it. It’s not the wrong answer but layering AI onto the current infrastructure of disparate systems will compound the problem to an unrecoverable state.
Understanding the telecom AI problem
Let’s say that I’m a telecom customer and I want a refund. I might go to the telecom’s website and interact with a chatbot that offers $5 off my next bill. The bot now sends this request to an API that was built 20 years ago. It breaks. Then I get routed to an agent. The agent now has to extract the information from the bot, and me (again) then manually alter a half-dozen systems to activate the refund. Where a large language model attempted to solve a problem, it created even more problems for both the company and the consumer. Meanwhile, all of the relevant information from that interaction is sent to different systems where it is parked. This is data at rest.
Now, should a telecom want to leverage AI for insights from that customer interaction, and others, the company first must ask an entire team of employees to dig through separate systems to locate the appropriate data (which now might be out of date or irrelevant), then feed that data into the system. It’s wildly inefficient and cost prohibitive. (An entire industry has fed off of this fractured establishment. Adtech and martech business suck billions from telecoms every year because we can’t take advantage of what’s sitting inside our very own systems of record.)
Solving the problem with billing (and AI)
The truth is, layering AI into telecom’s infrastructure is exactly the right thing to do. But where, and now? Billing is the key. But first, the infrastructure must be built or rebuilt to accommodate these modern tools and ensure the underlying problem isn’t exacerbated.
A modular, event-based architecture is precisely the foundation that telecom needs. In this system, every interaction from a customer, whether through a chatbot or with a human, creates an “event”. This event is then recorded, along with the data it produces, and shared with its appropriate system of record and any other system that needs to be aware, automatically. Layering AI on top of millions of events gives the machines exactly what they need–mountains of data that’s easily accessed, organized and extrapolated for insights.
Now the data is precise, current, and readily accessible. A human can ask a very simple request to create a complete profile of events–as specific as he or she wants it to be. This is data in motion. And it all comes down to billing–telecom’s single most important collector and funnel of customer information.
Consider all of the data and resulting insights that could come from one customer call about billing.
-
When did the customer call?
-
What was their issue?
-
What plan are they on?
-
What device do they use?
-
Where are they located?
-
What other services do they use?
-
How did we respond to their issue?
-
Are they now a churn risk?
-
What changes can we make to their account to solve multiple issues today and in the future?
-
What new products or services might they benefit from?
With the right infrastructure in place, layering AI with billing can help telecoms make more informed business decisions. We can easily create look-alike models and LTV models. We can be more thoughtful and strategic with new product development. We could even help solve one of telecom’s most fundamental data problems–a fractured base with millions of customers all on customized plans, custom SKUs, and individualized billing structures.
Conclusion
Telecom’s future is not just solving errors, but gaining insights. And the best place to find those insights is in our billing systems–one system of record that spans across customer care, marketing, and product development. A single view of the customer that is indisputable.
At Wavelo, we focus on building event-based, modular software infrastructures to help telecoms streamline operations, reduce costs and unlock opportunities–like applying rapidly advancing AI and large language models to grow revenue. Whether helping new service providers like DISH get a 15-year head start by building a new and modern base, or reconstructing billing systems for existing telecoms, we’re confident that this approach will revolutionize how telecoms service customers and grow their businesses today and far into the future.