- Executive summary
- AI’s growing role in customer service
- Customer care, transformed
- The soul of the machine?
- Hyper-personalisation of customer service
- Consumers still prefer the human touch
- Where to draw the line?
- Learnings
Executive summary
In a relatively short period of time, generative AI technology has found a wide range of applications across many different fields. Customer service, in particular, could significantly benefit from generative AI’s ability to understand and respond to complex requests in a human-like manner, enabling companies to provide their customers a more personalised and empathetic automated experience.
- Companies like Hume AI and Soul Machines have developed AI-powered chatbots that can recognise human emotions.
- 73% of consumers expect companies to understand their unique needs and expectations.
- 88% of respondents say that the experience a company provides is just as important as its products or services.
- “Synergistic advancements in data analysis and chatbot innovation have seen Gen AI elevate the customer experience in two key areas: personalisation and automation”, says Alex Rutter, managing director at Google Cloud.
- “Ultimately, personalisation isn’t possible without data, but companies must ensure it’s handled responsibly and used only in ways that customers approve”, says Jim Rudall, head of EMEA at email and marketing automation platform Intuit Mailchimp.
It appears increasingly likely that generative AI will play a greater role in customer service operations in the future. But is this really the right approach? Will AI ever be able to faithfully replicate every nuance of human interaction?
AI’s growing role in customer service
In the short period of time since ChatGPT was made public, numerous industries have been disrupted. Organisations of all types and sizes have wasted little time trying to figure out how it might boost their operations. Its unprecedented ability to understand and respond to complex inquiries in a manner that closely mimics human interaction has proven particularly useful in customer service. Companies in the sector have readily embraced it with the hope of handling larger volumes of customer queries, providing around-the-clock support, and lowering their response times.
As promising as generative AI technology arguably may seem, however, some questions remain about the true extent of its capabilities. Can it be trusted to properly interpret the full spectrum of human emotions and respond appropriately? Can it build meaningful, long-term relationships with customers in the same way a human agent might? And even if it can, should we let it? The human touch is precisely what makes a customer experience feel meaningful—what if we’re losing sight of that?
Customer care, transformed
Generative AI has the potential to revolutionise the way companies interact with their customers, but concerns about factual inaccuracies and bias remain.
Although customer care has a historic tendency to be overlooked in favour of profit, the idea of building deeper emotional connections with customers has gradually climbed the list of many companies’ priorities in recent years. Let’s be clear, though—this shift in mindset was not motivated entirely by altruism. An emotionally invested customer is, after all, also a loyal customer who will keep coming back, guaranteeing a steady stream of revenue.
The increased focus on customer care was accompanied by the digitalisation of the contact centre, which has transformed into something far removed from its humble analogue beginnings. Although the implementation of technologies like interactive voice response (IVR), robotic process automation, and chatbots did allow companies to expand their customer service capacity at minimal cost, they also came with noticeable limitations. We all understand the acute frustration of hitting a wall with a chatbot. In most cases, a canned response that does not address your query feels worse than receiving no help whatsoever.
The emergence of generative AI, on the other hand, has facilitated the creation of a new generation of chatbots that are capable of understanding more than just context and keywords. Indeed, they can also recognise the user’s emotional state, show empathy, and adjust their responses in real time based on the user’s evolving emotional needs.
While generative AI does represent a meaningful improvement over previous iterations of the technology, it is not without its own issues. The biggest red flag may be its tendency to produce factually incorrect information—a particularly serious concern in the context of customer care. There is also the risk of bias present in the data used to train the AI, which could cause it to inadvertently discriminate against certain segments of the population. The only way to mitigate these risks is to make sure a human agent reviews the content of each message before it’s sent to the customer. But wouldn’t that defeat the entire purpose of automation?
The soul of the machine?
Recent advances in generative AI have led to the development of the next generation of chatbots capable of understanding human emotions.
A number of companies out there are already working on expanding the capabilities of chatbots. New York-based research lab and technology company Hume AI recently announced the launch of its flagship product, Emphatic Voice Interface (EVI). Lauded by the company as “the first AI with emotional intelligence”, EVI uses an emphatic large language model (eLLM)—which is described as a combination of large language models (LLMs) and expression measurement models—to recognise emotions in a user’s voice and provide contextually appropriate responses that take into account the user’s emotional state. If a user sounds sad, for example, EVI will try to lift their spirits by offering words of encouragement or advice.
EVI employs the latest advances in semantic space theory—a data-driven, computational approach to understanding human emotions—to unearth the hidden meaning behind the pitch and tone of spoken language. “The main limitation of current AI systems is that they’re guided by superficial human ratings and instructions, which are error-prone and fail to tap into AI’s vast potential to come up with new ways to make people happy”, says Alan Cowen, founder of Hume AI. “By building AI that learns directly from proxies of human happiness, we’re effectively teaching it to reconstruct human preferences from first principles and then update that knowledge with every new person it talks to and every new application it’s embedded in”.
Another company aiming to help companies build stronger, more empathetic connections with their customers is San Francisco-based Soul Machines. It has developed what it calls ‘Digital People’—AI-powered virtual assistants that can interact with customers in a more natural and responsive manner. “The digital media landscape has changed tremendously in the last couple of decades”, says Greg Cross, co-founder of Soul Machines. “Brands are looking for new ways to connect. The big challenge for any company is how to build those brand connections”.
While traditional customer service chatbots typically rely on pre-scripted content to respond to customer queries, Digital People are completely autonomous and can react to the customer’s words and body language in real time. Each Digital Person is equipped with a ‘Digital Brain’, a proprietary technology that aims to simulate biological and neurological systems within the human body, rather than just mimicking the appearance of emotions. This enables them to produce more authentic, human-like responses and engage in more meaningful interactions with customers. They can also learn over time and adjust their responses to provide customers with a more personalised experience.
“AI is accelerating new methods for marketers to connect to their customers, and it will help them tailor their communications for better engagement and impact. This allows teams to focus human creativity towards the ‘why’ and ‘what’ of communication, whilst AI handles the ‘how’”.
Jim Rudall, head of EMEA at Intuit Mailchimp
Hyper-personalisation of customer service
While experts largely agree on generative AI’s ability to provide more personalised customer service, some are concerned that developing deeper bonds with AI could have unintended consequences.
Experts largely agree that generative AI can help companies provide their customers with more personalised experiences. “Synergistic advancements in data analysis and chatbot innovation have seen generative AI elevate the customer experience in two key areas: personalisation and automation”, explains Alex Rutter, managing director at Google Cloud. “On a customer-facing level, generative AI can unlock personalised shopping through AI-powered virtual agents. By answering customer queries with data-backed recommendations, chatbot agents can take the load off retail contact centres, ensuring customers receive seamless and efficient service”.
Jim Rudall, head of EMEA at email and marketing automation platform Intuit Mailchimp, is similarly convinced that generative AI will transform how companies engage with customers. “The future of email marketing and customer experience is evolving to be hyper-personalised”, he states. “AI is accelerating new methods for marketers to connect to their customers, and it will help them tailor their communications for better engagement and impact. This allows teams to focus human creativity towards the ‘why’ and ‘what’ of communication, whilst AI handles the ‘how’”.
However, Rudall also warns that companies need to proceed without compromising customer privacy. “Ultimately, personalisation isn’t possible without data, but companies must ensure it’s handled responsibly and used only in ways that customers approve”, he adds. While he agrees that generative AI can help companies cater to customers’ individual needs, Ori Bauer, chief executive officer of personalisation platform Dynamic Yield, believes that companies need to consider their approach with utmost care. “If you just let generative AI, or AI in general, run the show completely, it may come up with the wrong answers or with things that are not what you meant it to do”, he says.
Increased interaction with AI avatars could also lead humans to develop unintended—and potentially inappropriate—responses. “It could cause a lot of emotional distress with the people who interact with these systems and develop bonds”, explains Jackie Cheung, a McGill University professor and consultant with Microsoft Research. “There are lots of longer-term consequences that we haven’t really thought much about”. These thoughts are echoed by Sam Altman, chief executive of OpenAI. “I personally really have deep misgivings about this vision of the future where everyone is super close to AI friends, and not more so with their human friends”, he says, adding that users need to be aware they are interacting with a machine. “We named it ChatGPT and not a person’s name very intentionally”, adds Altman.
Consumers still prefer the human touch
A growing number of companies are integrating generative AI into their customer service operations, but consumers still prefer talking to a human agent over a machine.
It’s probably safe to say that today’s consumers are more demanding than ever. For example, we already knew in 2022 that 73% of consumers expect companies to understand their unique needs and expectations. Perhaps even more significantly, 88% of respondents say that the experience a company provides is just as important as its products or services. If they don’t get what they want, they will not hesitate to go somewhere else, with 52% of consumers saying they would switch to a competitor after a single negative experience, reveals Zendesk’s CX Trends 2023 survey.
In order to deliver the best possible customer experience, a growing number of companies are turning to generative AI. A 2022 Boston Consulting Group (BCG) survey of global customer service leaders revealed that 95% of respondents believe that at least some aspects of their customer service interactions will be handled by an AI bot within the next three years. Similarly, Gartner predicts that 80% of customer service and support organisations will use some form of generative AI to improve agent productivity and customer experience by 2025. BCG estimates that generative AI has the potential to increase productivity by anywhere between 30% and 50%.
So, how do consumers themselves feel about being served by an AI? While they are not categorically opposed to the idea, 43% of people would still prefer to receive customer service from a human if given a choice, according to research conducted by SurveyMonkey. A further 81% would be willing to wait a certain amount of time—anywhere between one and 11 minutes—in order to talk to a human customer service agent rather than being served by a bot, reveals a 2024 survey conducted by the customer experience platform Callvu. Interestingly, only 14% of respondents said they would rather skip the wait and talk to an AI immediately. When asked to compare their performance directly, respondents said that human agents outperformed their AI counterparts in 7 out of 10 critical customer service tasks.
“I personally really have deep misgivings about this vision of the future where everyone is super close to AI friends, and not more so with their human friends”.
Sam Altman, chief executive of OpenAI
Where to draw the line?
Despite the concerns surrounding the use of generative AI, it’s looking increasingly likely it will become an increasingly prominent aspect of customer service in the coming years. While humans will remain a fixture for the foreseeable future, further technological advancements will enable chatbots to resolve ever-more complex queries and deliver increasingly human-like responses.
The technology could become so advanced that it will enable companies to transition from a reactive approach to customer care to a more proactive one. Instead of trying to resolve problems only after they occur, generative AI-powered assistants will be able to detect early warning signs and reach out to customers, effectively nipping the issue in the bud. The need for human oversight will likely diminish gradually until humans are removed from the equation almost entirely, with every step of the typical customer journey being handled by AI.
It is, of course, easy to envision how this could go wrong. As customer service chatbots become increasingly human-like, people will grow to trust them more, reducing their inclination to question the information they provide. But what happens if that information is erroneous or even dangerous? Could advice from an AI assistant one day result in serious bodily harm or even the death of a customer? And if we don’t want to talk to a chatbot, will we be given a choice? Could we see “the right to talk to a human” be added to consumer protection laws right next to “the right to be forgotten”?
Learnings
The emergence of generative AI has had a transformative impact on the way companies interact with their customers. Thanks to their ability to understand complex inquiries and respond in a manner that closely mimics human interaction, AI-powered chatbots enable companies to provide faster, more efficient, and more personalised customer support than was previously possible.
- The potential benefits of generative AI customer service include 24/7 support, faster response times, and the ability to handle larger volumes of customer enquiries.
- The biggest limitation of generative AI is its tendency to produce factually incorrect information.
- Gartner predicts that 80% of customer service and support organisations will use some form of generative AI by 2025.
- 43% of people would still prefer to receive customer service from a human if given a choice.
- There are concerns that increased reliance on generative AI could reduce our ability to discern the reliability of information we are presented with.
While the benefits of generative AI are quite evident, there is something to be said for the human touch. The warmth you hear in another person’s voice, the sense that their concern is sincere, and feeling ‘understood’—these are the things that create true connections and foster loyalty between a brand and its customers. Are we prepared to sacrifice that just to get our problem resolved a little bit faster?
