Replacing human agents with AI chatbots could save the call centre industry up to $80bn in labour costs per year by 2026 and $240bn by 2031, according to a new report by Gartner which also predicts $1.99bn will be spent by the sector on AI this year alone.
The report predicts that the number of interactions between customers and a call centre handled by AI will go from 2% in 2022 to more than 15% by 2026, and double to 30% by 2031.
About 17 million contact centre agents are working around the world today with 95% of costs associated with a call centre due to staffing. Daniel O’Connell, VP analyst for Gartner and author of the report, says that organisations are both struggling to recruit and needing to reduce those costs.
“Conversational AI makes agents more efficient and effective, while also improving the customer experience,” he said. This is through added value to the human agent, such as providing analytics or pre-call data gathering, but also by replacing human agents.
Conversational AI can automate all or part of the customer interaction process, according to Gartner, this is done through both voice and digital channels, through voicebots or chatbots.
O’Connell says this new trend towards greater automation is expected to have “transformational benefits to customer service and support organisations”.
“While automating a full interaction – also known as call containment or deflection – corresponds to significant cost savings, there is also value in partial containment, such as automating the identification of a customer's name, policy number and reason for calling,” said O’Connell.
"Capturing this information using AI could reduce up to a third of the interaction time that would typically be supported by a human agent."
The technology is still maturing, but is improving all the time thanks to natural language models and libraries, allowing the chatbots to be more human-like in their approach to conversations.
Gartner says that due to the rapid development of AI technology it is still maturing and so has a fragmented vendor landscape with complex deployment solutions, which is resulting in a more measured adoption over the next two years, but that will speed up as the technology matures.
The delay in the rapid spread and uptake of conversational AI will be due to a lack of people with appropriate skills in analytics, knowledge graph and natural language processing, according to O’Connell, as it requires professional resources and specialist skills that are in high demand.
Even if the staffing is found to implement the technology in a large call centre, “once built, the conversational AI capabilities must be continuously supported, updated and maintained, resulting in additional costs".
An AI call centre agent will cost you up to $2,000
Implementing a large-scale complex AI deployment on the scale required for a large call centre would take multiple years to implement. This is because call flows need to be built out and existing call flows fine-tuned to work with the automated solutions.
Gartner estimates integration pricing of $1,000 to $1,500 per conversational AI agent, though some organisations cite costs of up to $2,000 per agent. That is still considerably less than the average cost of employing a human agent, which recruitment service Reed puts at £22,630 per year.
The price tag for an AI agent means that the early adoption is being led by companies that have 2,500 or more agents and a budget to invest in the new technology.
“Assuming a baseline of professional services cost of $1,200 per agent (one time), we project efficiency benefits of 4% per year over the course of the forecast,” the report claims.
“This aggressive efficiency gain accrues from more professionals joining an attractive high-growth market, maturing (simpler) conversational AI platforms and experiences (best practices) learned across the market. Hence, per agent, professional services costs are projected to go down to $978 in 2026 and $798 in 2031.”
There will be different levels of adoption and automation. At a basic level, before full integration, AI could be used to identify the customer, gather their name, account number and reason for the call which could then be passed on to a human agent, freeing up the human’s time to take more calls.
The biggest challenge is how happy consumers will be to engage with artificial intelligence agents. A 2020 study by Zendesk found that about 40% of respondents to a survey in the UK were happy to engage with AI rather than a human if it got them a faster or more accurate result.
Tech Monitor is hosting a roundtable in association with Intel vPro on how to integrate security into operations. For more information, visit NSMG.live.