• Subtonomy

How to boost chatbot performance in the 5G era

Updated: Apr 11


Robot dance

CSPs love them and have spent millions on them, but customers frequently aren’t so impressed. Fredrik Edwall looks at what’s going wrong with bots and how we can fix it.


Back in 2016 bots were spoken about by the great and the good of the tech industry as the next big disruptive technology. Microsoft’s Satya Nadella commented: “Pretty much everyone today who's building applications, whether they be desktop apps, or mobile apps, or websites, will build bots as the new interface.” He wasn’t alone in his enthusiasm for bots. Gartner predicted that “the average person will have more conversations with bots than with their spouse” and Oracle forecast that 8 out of 10 business would be using chatbots by 2020.


CSPs became big bot believers and since 2016 they’ve invested millions in their AI-driven vision of customer service. But fast forward to 2021 and disillusion had set in.


Customers were nowhere near as enthusiastic about bots as CSPs, and unimpressed by all the money being lavished on them. This reversal of fortunes saw Gartner revise its opinion and put chatbots at the bottom of its hype cycle for Natural Language Technologies (July 2021). Omnisperience characterized them as “nice but dim”. And, rather damningly, research by Userlike found that 60% of customers said they’d rather queue for an agent than receive immediate service from a bot.


So what’s the deal with bots? Are they, as some research firms have called them, a “failed revolution”? Are they doomed to disappoint? Or can they still live up to the exciting potential that grabbed the attention of so many?

After the 2016 hype, reality quickly set in. CSPs realized that early chatbots struggled with natural language features such as accents and nuances, couldn’t understand customer intent, weren’t empathetic like human agents and didn’t have access to enough data to handle complex requests. In the five years since, much work has been done to address these shortfalls and create more performant and intelligent bots. Chatbots now incorporate conversational AI, semantic analysis and natural language processing, which enable them to provide both more accurate and more personalized support.

The Covid-19 pandemic of 2020-21 was a major boost for bots. Chatbots were deployed to fill the gaps as companies moved to remote working, staff were off sick and the volume of inquiries increased. Some news reports even began to talk about bots as a call centre industry killer, as they stepped into the gap to provide 24x7 support.


But it’s not yet time to write off the call center. The truth, as always, is a little more mundane. The way bots are being used has changed, augmenting the call center rather than replacing it within an increasingly complex multichannel mix. Standalone bots handle simple inquiries and triage customers before they’re transferred to human agents – meaning agents can concentrate on the most complex inquiries or the most frustrated customers. They provide a quick, digital alternative for customers that don’t want to talk to agents. But bots are also being deployed within call centers, acting as a smart assistant to customer service agents – providing key information about customers to enhance their performance.


As Innovation and Data Science Director Alan Linter explained at the launch of Capita’s Assisted Customer Conversations tool: “By providing real-time rather than historic analytics, coupled with AI-driven advisor assistance, it will enable our advisors not just to be more responsive, but to respond better”.

Another key trend is moving beyond bots that provide support based on static, historic data that enables them to answer ‘frequently asked questions’ to feeding bots with the dynamic, real time data that enables them to solve customers’ current and unique problems.


In telecoms the most common inquiries to call centers relate to bills and service quality issues such as outages and slow service. In both cases bots and customer service agents need access to real time data in order to efficiently handle customer inquiries.


Led sign saying Data has a better idea

This is where Subtonomy helps. A common misunderstanding we hear all the time is that bots replace the need for our platform. In fact the reverse is true. Our platform is a key enabler of bots, customer service agents and self-service tools - simultaneously providing realtime network performance data to all care channels and ensuring they’re all working from the same, accurate and up-to-date data. Feeding bots with such data not only enables the bot to be more performant and therefore provide more accurate, personalized service to CSPs’ customers, but also ensures that all customer service channels are in-sync. This means that if a customer needs to be transferred from self-service or chat bot to a call center agent, the agent they speak to is working from exactly the same data – driving a seamless, consistent experience.


Whatever the combination of self-service, AI bot, or human service agent CSPs deploy and customers choose during their support journey, Subtonomy’s real time network performance data provides the insight required to resolve their inquiry faster. We help our customers create the type of support experience customers increasingly expect in the 5G era by putting realtime network performance data at CSRs’ fingertips to help them resolve technical support inquiries faster, and by feeding chatbots with the same data we make them ‘brainer’ and able to handle a greater range of inquiries. Read more about our Customer care solutions