When asked whether AI will supplant human workers in the customer service sector, ChatGPT typically offers a balanced perspective, suggesting that the two will likely coexist.
However, human sentiment is often less sanguine.
K Krithivasan, chief executive of Tata Consultancy Services, posited to the Financial Times last year that AI advancements may soon lead to a “minimal need” for Asian call centers.
Furthermore, Gartner, a business and technology research firm, forecasts that AI will autonomously resolve 80% of routine customer service inquiries by 2029.
The current focus is heavily centered around “AI agents,” a term denoting AI systems with enhanced autonomy and decision-making capabilities.
These agents could significantly augment existing “rule-based chatbots,” which are limited to answering a predefined set of questions.
A recent personal experience with parcel delivery firm Evri’s chatbot highlights the current state of affairs with non-AI systems.
When a parcel failed to arrive, Ezra, the chatbot, offered to “get this resolved straight away.”
After requesting and receiving the tracking reference, Ezra stated that the parcel had been delivered.
Upon requesting proof of delivery, the system provided a photo of the package… at the wrong address. No further options were available to advance the conversation after the presentation of this “evidence.”
In response, Evri told the BBC that they are investing £57m to improve their service.
“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,” the company stated.
“Our data confirms the vast majority of people get the answers they need from our chat facility, first time, within seconds. We’re always reviewing feedback to ensure our services are as helpful as possible, and we continue to make enhancements on a rolling basis.”
Conversely, rival DPD had to disable its AI chatbot after it criticised the company and swore at users.
Balancing brand representation with genuine customer assistance is a challenge for businesses as they transition to AI.
Gartner reports that 85% of customer service leaders are actively exploring, piloting, or deploying AI chatbots. However, only 20% of these projects fully meet expectations.
“You can have a much more natural conversation with AI,” says Gartner analyst Emily Potosky.
“But the downside is the chatbot could hallucinate, it could give you out-of-date information, or tell you completely the wrong thing. For parcel delivery I would say rules-based agents are great because there are only so many permutations of questions about someone’s package.”
Resources and cost savings are key drivers for businesses considering the shift from human to AI customer service. However, Ms. Potosky notes that AI is not necessarily a cheaper alternative.
“This is a very expensive technology,” she says.
The first step for any business contemplating replacing human agents with AI is to ensure the availability of extensive training data.
“There’s this idea that knowledge management becomes less important because generative AI can solve the fact that their knowledge is not particularly well organised, but actually the opposite is the case,” adds Ms. Potosky.
“Knowledge management is more important when deploying generative AI.”
Joe Inzerillo, chief digital officer at Salesforce, told the BBC that call centers provide fertile training grounds for AI, particularly those in low-cost regions such as the Philippines and India.
This is due to the comprehensive staff training provided, which AI systems can also 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 says.
Salesforce’s AI-powered customer service platform, AgentForce, is currently used by organizations ranging from Formula 1 to Prudential, OpenTable, and Reddit.
Mr. Inzerillo notes that initial trials of the platform revealed valuable insights into making the AI more human-like.
“While a human might say ‘sorry to hear that’, the agent just opened a ticket,” says Mr. Inzerillo.
As a result, the AI was trained to demonstrate empathy, particularly when addressing customer issues.
Salesforce also discovered that restricting the agent from discussing competitors posed problems.
“This backfired when customers asked legitimate questions about integrating Microsoft Teams with Salesforce,” says Mr. Inzerillo. “The agent refused to help because Microsoft appeared on our competitor list.”
The firm subsequently revised this policy.
Salesforce has ambitious plans for the continued deployment of its AI agents and reports high customer satisfaction. The company also claims that 94% of customers choose to interact with AI agents when given the option.
“We’ve seen customer satisfaction rates that are in excess of what people get with humans – then AI can unlock the next level of customer service,” says Mr. Inzerillo.
The platform has also resulted in a $100 million reduction in customer service costs, although Mr. Inzerillo downplayed reports suggesting this has led to 4,000 job cuts.
“A very large percentage of those people got redeployed in other areas around customer service.”
Fiona Coleman leads QStory, a company utilizing AI to provide human call center workers with greater flexibility in their shift patterns. Their clients include eBay and NatWest.
While acknowledging the value of AI in improving working conditions, she doubts the technology can entirely replace human agents.
“There are times where I don’t want to have a digital engagement, and I want to speak to a human,” she says.
“Let’s see what it looks like in five years’ time – whether an AI can do a mortgage application, or talk about a debt problem. Let’s see whether the AI has got empathetic enough.”
Indeed, a backlash against the use of AI in customer service may already be brewing.
Proposed US legislation aimed at relocating offshore call centers back to America also requires businesses to disclose the use of AI and transfer callers to a human agent upon request.
Gartner also predicts that by 2028, the EU may mandate a “right to talk to a human” as part of its consumer protection regulations.
Richard Stott had “unanimous support” from people agreeing AI interviews were “not a good idea”.
Your social media feed is being taken over by AI video slop. But for now, there’s still one big red flag that can help you spot the fakes.
The president pardoned Changpeng Zhao, the co-founder of the world’s largest crypto exchange, in October.
Researchers say this new method rethinks how AI systems are wired at their most fundamental level.
Fresh from $5 trillion valuation, global chips giant Nvidia has signed deals with LG, Hyundai and Samsung