“Enterprise AI” is the Real Future of Customer Service
AI is transforming customer service faster than most could have imagined just a year ago. Gartner predicts that by 2029, AI will autonomously resolve 80% of common customer issues, a huge impact on the CX industry that for many years has had few other options than to try and reduce their costs by outsourcing. Tata Consultancy Services’ CEO earlier this year suggested that AI impact could make many call centers in Asia largely unnecessary.
Now at this point, sourcing to low-cost areas is not likely to disappear, but the amount of automation options AI brings, will reduce the need for employees, including those low-cost employees. Perhaps this could even lead to companies rethinking their cost strategies for customer service due to AI.
But while headlines often focus on systems like ChatGPT or Gemini, the most effective breakthroughs within CX are happening in a different space altogether, one we at Nexcom have labeled Enterprise AI – an AI that is company-specific, trained on a business’s own data and operating processes.
From general intelligence to company expertise
Generic AI assistants are impressive for their range. They can summarize research, translate text, draft an email for a sales pitch or suggest an agenda for the company meeting. But when a customer’s parcel goes missing or a hotel guest’s booking fails, general intelligence doesn’t help much.
A generic AI will know what a refund is — but not how your company processes one. It can talk about delivery delays and how frustrating they are, but it cannot follow your internal escalation path and do anything about it, and that’s why many AI implementations have either straight out failed, or at least not brought the needed benefits. A Gartner study was released where just 20% of AI initiatives among participants were successful in Customer Service and please remember that 75% of all errors that happen in a call center are outside the employees’ control. IT is not enough to believe that knowledge management is key and think you are all done when you’ve “uploaded” the knowledgebase.
Enterprise AI, on the other hand, is trained to understand your world before it engages in supporting customers. It knows your workflows, service logic and processes in order to be effective. Then it can resolve real-world issues - and it can do it in your brand’s tone of voice.
Our experience is that in order to succeed, it’s not enough to have good data discipline. There’s this idea that generative AI can solve the fact that knowledge is not particularly well organized and magically get things right because it is an AI. That is wrong and the opposite is actually the case. Knowledge and insight into how the data is generated is more important when deploying generative AI. Knowing where to look for the right data and what data to discard is crucial.
Where technology meets operational insight
Creating effective Enterprise AI is about more than algorithms. It demands insight into how customer service really works - from receiving an inquiry to empathy and tone. It’s all about that all important customer experience – you can’t afford to lose customers due to AI now, can you?
While here at Nexcom, I’ve seen first-hand how important that is. Call centers, where thousands of conversations take place every day, can be ideal training grounds for an AI – providing you information on how the data was generated in the first place. They may contain the documentation, context and nuance that help AI truly understand how to handle a customer’s issue - but remember, they also have lots of errors and wrong answers, and it is crucial that an AI knows the difference.
Early in our own development process, we learned that technical accuracy alone wasn’t enough. An example was our voice AI could handle calls flawlessly — but it sounded like what it was, a robot. It didn’t pause, reassure, or empathize the way a skilled agent would. We named it the “hm hm” factor – small noises that we humans use to show we are listening, and we adjust our tone of voice depending on how we are spoken to. People expect that and the absence of it results in a bland customer experience. Hence this was incorporated in the next version of VIBE.
Once we trained the system to understand tone, pacing and emotional context, satisfaction levels rose sharply — and we can see now that customers chose first to interact with the AI when given the option. As consumers we choose convenience, and now that voice AI can be a good experience, consumers start choosing it first as it is more convenient.
That’s the essence of Enterprise AI: answering correctly AND doing so in a way that feels helpful.
Built for experience, not just efficiency
Enterprise AI is not just about replacing people. It’s about extending a company’s best knowledge and service to every customer interaction — 24/7, without sacrificing the customer experience.
In that sense, the future of AI in customer service won’t belong to the most powerful model, but to the most purpose-built one. The Enterprise AI that understands your customers because it understands you.
Welcome to Enterprise AI!