AI & Call Centres: Is The End Nigh?

AI in Customer Service: Finding the Balance Between Bots and Humans

We're constantly hearing about AI working alongside people, yet at the same time, industry leaders are hinting that traditional call centres could soon be a thing of the past. It really makes you wonder where human interaction fits into this rapidly evolving technological landscape.

Just look at what K. Krithivasan, the boss at Tata Consultancy Services, recently told the Financial Times. He suggested that AI might lead to "minimal need" for call centres, especially those located in places like Asia. This echoes a broader trend where the AI Wave Shifts to Global South, impacting various industries.

The tech research gurus at Gartner are forecasting that a staggering 80% of everyday customer service queries will be handled entirely by AI by 2029.

That's a significant shift we all need to get our heads around, isn't it?

From Basic Chatbots to Clever AI Agents

We're well past the days of simple, rule-based chatbots, the kind we've all encountered that can only answer a very specific set of questions. Now, the conversation has moved on to "AI agents." These are far more sophisticated, designed to make decisions and operate with a good deal more autonomy. They represent a huge leap forward in how companies can interact with their customers, raising questions like Will AI Agents Steal Your Job Or Help You Do It Better?.

I recently had a rather frustrating experience with Evri's chatbot, Ezra, which perfectly illustrated the limitations of these older, non-AI systems. My parcel hadn't arrived, and Ezra confidently promised to "get this resolved straight away." After I provided the tracking reference, it informed me my parcel had been delivered. When I asked for proof, it promptly showed me a perfectly clear photo of the package on someone else's doorstep.

The really annoying part? There was no option to take the conversation any further after this "evidence" was presented.

Now, credit where credit's due, Evri has told the BBC they're investing £57 million into improving their service. They stated, "Our intelligent chat facility uses tracking data to suggest the most helpful responses and ensure the customer's parcel is delivered as soon as possible, if this has not happened as scheduled."

They also claimed that "the vast majority of people get the answers they need from our chat facility, first time, within seconds." While that's probably true for straightforward queries, my little saga highlights there's still a gap when things go a bit wrong.

On the other hand, we've seen things go spectacularly awry with less constrained AI. Rival firm DPD actually had to pull the plug on their AI chatbot after it started slagging off the company and even swearing at users. Definitely not the kind of "customer service" anyone's looking for!

"You can have a much more natural conversation with AI... But the downside is the chatbot could hallucinate, it could give you out-of-date information, or tell you completely the wrong thing." - Emily Potosky, Gartner analyst

Finding that sweet spot between staying on brand and genuinely helping customers is proving a rather tricky tightrope walk for businesses experimenting with AI.

Big Hopes, Tricky Realities: The AI Customer Service Challenge

It's pretty clear why so many companies are keen on AI chatbots, isn't it? Gartner found that 85% of customer service leaders are either looking into them, piloting them, or actually deploying them. However, the same research also revealed that only 20% of these projects are truly meeting expectations. That's a significant difference between what people hope for and what they're actually getting.

Emily Potosky, a Gartner analyst, makes a really important point: while AI can offer a much more natural chat, it also comes with risks.

"The downside is the chatbot could hallucinate, it could give you out-of-date information, or tell you completely the wrong thing," she explains.

For something like parcel delivery, where questions often have clear, limited answers, those older, rule-based systems might actually be a better fit.

Many businesses are viewing AI as a way to cut costs, hoping to replace human agents. But Ms Potosky warns that AI isn't necessarily cheaper in the long run. "This is a very expensive technology," she notes. For AI to work properly, companies need loads of well-organised training data. In fact, she argues that knowledge management becomes even more crucial when you're rolling out generative AI, not less. It really highlights the importance of getting your data house in order. For more insights on this, Gartner's report on "The Future of Customer Service" provides further details^. The Future of Customer Service

Training the Bots: Learning from the Best

Joe Inzerillo, the chief digital officer at software giant Salesforce, believes that call centres, especially those in places like the Philippines and India, are absolutely brilliant training grounds for AI. Why? Because they often have a goldmine of documented processes and staff training materials that AI can learn from. "You have a huge amount of documentation, and that's all really great stuff for the AI to have when it is going to take over that first line of defence," he told the BBC. This is particularly relevant given India's AI Future: New Ethics Boards which will likely influence how AI is trained and deployed in such regions.

Salesforce's AI-powered platform, AgentForce, is already being used by big names like Formula 1, Prudential, OpenTable, and Reddit. During its development, Salesforce picked up some valuable lessons about making the AI feel more "human-like." Initially, the AI would just "open a ticket" when a person might say "sorry to hear that." So, they trained the AI to show a bit more empathy, particularly when customers were expressing a problem.

Another interesting lesson came up around competitor discussions. The AI was originally programmed to avoid talking about rivals, which completely backfired when customers asked perfectly reasonable questions about integrating, say, Microsoft Teams with Salesforce. The AI just clammed up. Salesforce quickly realised this rigid rule was counterproductive and changed it.

Salesforce has big plans for its AI agents and claims they're a huge hit, with 94% of customers choosing to interact with AI when given the option. Mr Inzerillo even suggests they've seen "customer satisfaction rates that are in excess of what people get with humans."

He also mentioned a $100 million cut in customer service costs, though he was quick to clarify that this didn't mean 4,000 job losses, but rather people being moved into other areas within customer service.

The Lasting Need for a Human Touch

Despite all these amazing AI advancements, many people, like Fiona Coleman from QStory, a firm that uses AI to improve human call centre shift patterns, believe there will always be a need for human interaction. Her company works with big names like eBay and NatWest.

"There are times where I don't want to have a digital engagement, and I want to speak to a human." - Fiona Coleman, QStory

It's a sentiment many of us can relate to, isn't it? When you're dealing with genuinely complex issues like a mortgage application or problems with debt, empathy and a nuanced understanding are absolutely vital. "Let's see whether the AI has got empathetic enough," she muses, suggesting we might be a good five years away from AI handling such sensitive conversations.

In fact, there's even a bit of a backlash brewing. In the US, proposed legislation to bring off-shore call centres back home also includes rules for businesses to disclose when AI is being used and to transfer callers to a human if requested.

Gartner also predicts that by 2028, the EU might even enshrine a "right to talk to a human" into its consumer protection rules.

So, while AI is undoubtedly changing customer service, it seems the human element, both in its need for empathy and its right to choose, isn't going anywhere just yet.

The future probably isn't about completely replacing people, but rather finding the right balance where AI handles the routine, and humans excel at those truly complex and emotionally charged conversations.