Zendesk recently participated in The Economist Impact panel, which brought technology industry leaders and analysts together to explore how AI is transforming customer service. The conversation was moderated by The Economist’s US Technology Editor, Guy Scriven, and in addition to Zendesk’s Chief Technology Officer, Adrian McDermott, they were joined by:
- Jessica Moulton, Head of McKinsey’s Consumer Packaged Goods Practice in Europe
- Mohamed Abdelsadek, Executive Vice President of Data, Insights and Analytics at Mastercard
- Kate Leggett, Vice President and Principal Analyst at Forrester Research
Read on for highlights from the panel.
How are advances in technology and evolving consumer expectations transforming customer service? What are the impacts of AI and personalization?
Customer service has become one of the top ways to differentiate from the competition. Adrian cited the 2023 Zendesk CX Trends Report which asks consumers about the likelihood of switching to a competitor after one bad customer experience. Every year, this number climbs as expectations continue to rise proving how a great customer experience now makes all the difference.
In early 2023, Zendesk stats showed that more than half of customers will switch to a competitor after one bad experience and nearly three-fourths will after two bad experiences. “We see this impact globally – there’s no question across the world customer service is becoming more and more critical to earn and keep customers,” said McDermott. “A big part of what makes a great customer experience is fast, effective, and personalized support. Technology, especially AI, is a critical part in creating that positive experience – not just for the customer, but for every role in support like for agents and admins too.”
Leggett echoed this sentiment, noting Forrester Research data that shows two-thirds of interactions that a customer has with a brand are with customer service (i.e. pre-purchase, onboarding, post-purchase). “Great customer service impacts retention, value over time, ability to advocate and be a reference for a brand,” she said. “These levels of loyalty impact topline revenue in dozens of industries and geographies.”
McDermott continued to say that AI has already improved CX in many ways but it’s still just the beginning. AI is going to absolutely transform CX dramatically over the next decade as there is a lot of manual and repetitive work in CX and agent time spent digging for information – AI is perfect to handle these tasks.
Will generative AI be transformative for customer service, and if so, how?
While AI has been improving aspects of CX for quite some time, the impact has really been on simple support needs. McDermott explained how generative AI opens the door for AI to effectively assist in complicated inquiries due to its ability to understand much more about the customer’s needs. This ranges from tasks like understanding a customer’s exact intent (for ex: make a return, update an account, etc.); parsing through company information to find the right answer; and engaging with a customer personally and conversationally.
“Generative AI will change the nature of support. No longer will people accept ten blue link responses (like you get in a Google search today), they will just want the direct answer,” he said. “Gen AI also really opened our eyes to nothing being off limits for AI – for example, we used to think reasoning and creativity were the unique domain of the human species, which is no longer the case.”
Looking at this trend from a global lens, there is a massive opportunity for generative AI to be especially useful in providing support in more regional languages. Also in Latin America, generative AI can make conversational support even easier, where consumers are more accustomed to messaging and expecting it from businesses.
“The agent role is changing from creators of content to supervisors of CX.” – Adrian McDermott, Zendesk CTO
Leggett added that AI will transform agents into “super agents.” She explained that having all the insights, content, knowledge pushed to agents proactively will allow them to better serve the customer. This makes room for agents to bring their full attention and empathy to the interaction. Additionally, Leggett noted the trend that there is an increased need for agents who are better skilled and have the empathy required to emotionally connect with a customer. In fact, she mentioned that many brands have shifted the profiles of agents that they hire.
How can businesses strike a balance between AI-driven automation and human interaction to enhance the customer experience?
With today’s challenging economy and increased competition, most leaders agree that controlling costs is a top priority. All businesses are reevaluating spending and exploring how to do more with less – which is where AI is especially appealing. At Zendesk, McDermott explained that the company has been building towards a future where eventually all service will be AI first.
“We’ve seen technology change the nature of employment time and time again since the Stone Age. That’s the whole point of technology and tools – to make tasks easier and more efficient for humans,” McDermott said.
Humans are then redeployed in elevated ways to perform more supervisory functions and roles. One very critical supervisory function humans will play as AI extends is ensure it’s safe, trustworthy, secure and ethical. “AI doesn’t have a moral compass – it doesn’t care if what it’s doing is fair, accurate, and unbiased,” he added. In addition to supervising AI to be ethical and responsible, there are also some specific customer experience risks that should be mitigated with humans supervising AI.
“In the near-term, we will still have human agents required for escalation and connecting directly with companies’ highest priority customers,” Moulton added. “Generative AI will supercharge many of these roles, but human touch will be needed for a long time.”
Adrian agreed, saying, “Humans like to talk to humans, especially when in an emergency situation.” Which brings us to the question: to automate or not to automate. “Models know a lot,” he said, “but still can’t reason at quite the same level of an agent which is why humans are critical to still be involved and supervise bots as needed.”
For more Zendesk POV on AI, visit our Newsroom.
