Ask ChatGPT if artificial intelligence will replace humans in customer service, and he’ll give you a diplomatic answer. Its gist will boil down to the phrase: “They will work side by side.”
If you ask humans, the answer is not so optimistic.
Last year, the CEO of Indian technology company Tata Consultancy Services, K. Kritivasan, told the Financial Times that AI could soon make the need for call centres in Asia “minimal”.
Meanwhile, business and technology research firm Gartner predicts that by 2029, artificial intelligence will solve 80 per cent of typical customer service queries on its own.
There’s a lot of hype around “AI agents” at the moment. The term refers to artificial intelligence systems capable of tracking, and when I introduced it, I said the parcel had already been delivered.
I could ask for proof of delivery, and when I did, the bot showed me a picture of the parcel… at someone else’s door. After this “proof” it was impossible to continue the conversation – no options he offered.
In response, Evry told the BBC it was investing £57 million to improve the service.
“Our smart chat uses tracking data to offer the most helpful responses and ensure a parcel is delivered as quickly as possible if it hasn’t been delivered on schedule,” the statement said.
“Our data confirms that most users get the right answer the first time – in seconds. We are constantly analysing feedback to make our services AI-powered, but only 20 per cent of these projects fully meet expectations.
“With AI, you can have a much more natural conversation,” says Gartner analyst Emily Potoski.
“But there’s a downside: chatbots can have ‘hallucinations,’ meaning they can make up information, provide outdated data or completely wrong answers. For package delivery, I would say that rule-based agents are a great solution because the number of options for questions about the package is limited.”
One of the main reasons companies are considering moving from human support to AI is resources and finances. But as Ms Potoski points out, we shouldn’t automatically assume that AI will be cheaper than human agents.
“It’s a very expensive technology,” she emphasises.
The first step for any business that wants to replace AI experts.
“Knowledge management becomes even more important when implementing AI.”
Joe Inzerillo, chief digital officer at Salesforce, told the BBC that call centres are fertile ground for AI training, especially those that have been moved to low-wage regions such as the Philippines and India.
This is because a lot of staff training has already been done there – and AI can also learn from that material.
“There is a huge amount of documentation and it’s a very useful resource for AI when it has to take on the role of the first line of defence,” says Inzerillo.
Salesforce’s AI-powered customer service platform AgentForce is already being used by clients such as Formula One, insurance company Prudential, restaurant booking site Open Table and social networking site Reddit.
Inzerillo says that during testing I said: “‘I’m sorry to hear that’ and the agent just opened the application,” he explains.
So AI is being taught to show more compassion, especially when a customer is facing a problem.
Salesforce has also found that forbidding an agent from mentioning competitors creates difficulties.
“It backfired on us when customers asked perfectly logical questions about integrating Microsoft Teams with Salesforce,” says Inzerillo. – The agent refused to help because Microsoft was listed as a competitor.”
The company has since revised that strict rule.
Salesforce has ambitious plans to further roll out its AI-based agents and currently claims they are proving popular with customers. It says the vast majority of users – 94 per cent – choose to interact with AI agents when they have the option.
“We c. jobs.
“A very large proportion of those staff have been redeployed to other customer service roles,” he added.
Photo by Fiona Coleman
Fiona Coleman heads QStory, a company that uses artificial intelligence to give call centre staff more flexibility in their work schedules. Her clients include eBay and NatWest.
While she recognises that AI can improve working conditions, Fiona doubts that the technology can completely replace humans.
“Sometimes I don’t want digital interaction – I want to talk to a human,” she says.
“We’ll see how things look in five years’ time – whether AI can arrange a mortgage or talk about a debt problem. We’ll see if it becomes empathic enough.”
However, the use of AI in customer service may already be facing resistance.
У