Does artificial intelligence mean the end of call centers?

Ask ChatGPT whether AI will replace humans in the customer service industry, and he will give a diplomatic answer, summed up as “they will work side by side.”

But humans are not that optimistic.

Last year, K Krithivasan, CEO of Indian technology company Tata Consultancy Services, told the Financial Times that artificial intelligence could soon mean there are… “Minimum need” for call centers in Asia.

Meanwhile, artificial intelligence will solve the problem independently 80% of common customer service issues By 2029, business and technology research firm Gartner predicts.

There is currently a lot of hype around “AI agents”. This is the term given to artificial intelligence systems that can operate more autonomously and make decisions.

They could motivate existing non-AI chatbots, known as “rules-based chatbots,” which can only answer a specific list of questions.

My recent experience with Evri’s chatbot for package delivery illustrates the current non-AI state of the game.

My package didn’t arrive, and Ezra (the name of the chatbot) offered to “resolve this issue immediately.”

I asked for a tracking reference, and after I wrote that down, it told me my package had been delivered.

I could ask for proof of delivery, and when I did it showed me a photo of the package…at the wrong front door. There was no option to continue the conversation after this “evidence” came to light.

In response, Every told the BBC it was investing £57m to improve the service further.

“Our intelligent chat tool uses tracking data to suggest the most helpful responses and ensure a customer’s package is delivered as soon as possible, if it doesn’t happen as scheduled,” she says.

“Our data confirms that the vast majority of people get the answers they need from our chat facility, first time, within seconds. We always review feedback to ensure our services are as useful as possible, and we continue to make improvements on a regular basis.”

On the other hand, competing parcel delivery company DPD was forced to disable its less-compliant chatbot after the company criticized and cursed users.

Striking the right balance between being on-brand and genuinely helping customers is a difficult thing for companies to navigate as they transition to AI.

some 85% of customer service leaders They are exploring, piloting or deploying AI-powered chatbots, according to Gartner. But it also found that only 20% of these projects fully met expectations.

“You can have a much more natural conversation with AI,” says Emily Potosky, an analyst at Garner.

“But the downside is that the chatbot could hallucinate, or it could give you outdated information, or tell you something that’s completely wrong. For package delivery, I would say rules-based agents are great because there are only so many permutations of questions about someone’s package.”

Resources and money are among the main reasons why companies may consider moving from human customer service to AI-based customer service. But Ms. Potoski points out that it is not a given that AI will be cheaper than human agents.

“It’s a very expensive technology,” she says.

The first thing any company that wants to replace humans with AI should do is make sure it has extensive training data.

“There is an idea that knowledge management becomes less important because generative AI can solve the fact that their knowledge is not well structured, but in fact the opposite is true,” Ms. Potoski adds.

“Knowledge management is even more important when deploying generative AI.”

Joe Inzerillo, chief digital officer at software giant Salesforce, tells the BBC that contact centers provide fertile ground for AI training, especially those that have been relocated to low-cost regions such as the Philippines and India.

This is because a lot of employee training will be done, which the AI ​​can also learn from.

“You have a huge amount of documentation, and those are all really great things for AI to have when it takes on the first line of defense,” he says.

Salesforce’s AI-powered customer service platform, AgentForce, is currently used by a range of clients from Formula 1, to insurance company Prudential, restaurant booking site Open Table, and social media site Reddit.

When Salesforce first experimented with the platform, Inzerillo says, it learned some valuable lessons about how to make AI seem more human-like.

“While a human might say, ‘Sorry to hear that,’ the agent just opened a ticket,” Inzerillo says.

So AI has been trained to show more empathy, especially when a customer has a problem.

Salesforce also found that not letting an agent talk about competitors was a problem.

“This backfired when customers asked legitimate questions about integrating Microsoft Teams with Salesforce,” Inzerillo says. “The agent refused to help because Microsoft appeared on our list of competitors.”

The company later replaced that strict rule.

Salesforce has ambitious plans to continue rolling out AI agents, and so far claims they are a big hit with its customers. It also says that the vast majority of customers, 94%, choose to interact with AI agents when given the option.

“We’ve seen customer satisfaction rates exceed what people get with humans — so AI can unlock the next level of customer service,” says Inzerillo.

It also means the company has cut customer service costs by $100 million, but he was keen to play down recent headlines suggesting this led to 4,000 job cuts.

“A very large percentage of those people have been redeployed into other customer service areas.”

Fiona Coleman runs QStory, a company that uses artificial intelligence to provide human call center workers with more flexibility in their shift patterns. Its clients include eBay and NatWest.

While she sees the value of AI in improving working conditions, she is not sure the technology can completely replace humans.

“There are times when I don’t want to engage digitally, and I want to talk to a human,” she says.

“Let’s see what it looks like in five years – whether AI can make a mortgage application, or talk about a debt problem. Let’s see if AI has become empathetic enough.”

In fact, the use of AI in customer service could already face backlash.

Legislation currently proposed in the United States to move offshore call centers to America also requires companies to disclose the use of artificial intelligence, and turn the caller into a human if asked to do so.

Meanwhile, Gartner predicts that by 2028 in the European Union May be delegated The so-called “right to talk to a human being” as part of consumer protection rules.