The advent of ChatGPT seems to represent a quantum leap in the capacity of artificial intelligence (AI) to reshape the world we live in. Things are moving quickly, and organisations must begin now to think about how they can harness the power of AI to improve their performance beyond what basic process automation has long made possible.

Customer service is especially ripe for the kind of improvements that AI can bring. Whether for providers of consumer goods or services ranging from healthcare to insurance and financial products—customer service operations have long been plagued by chronic inefficiencies. AI can help to reverse that.

However, it’s vital for organisations to bear in mind that people have always been and always will be the indispensable core of customer service delivery. New technologies like ChatGPT and other advanced AI applications must enhance—not replace—human customer service agents.

Don’t make the same mistake twice

Years ago, the automation of incoming call distribution and staff scheduling processes improved customer service efficiency, but it didn’t improve agents’ ability to satisfy customers once they were connected. Fast forward to today and this is where AI comes in. In a fraction of the time it takes for a human, AI can access and process massive quantities of relevant data and predict best solutions to customer issues.

However, organisations must not make the same mistake they made with early chatbots, which were falsely presumed to be a simple replacement for less reliable and more costly human agents. For years, chatbots have supported with simple, transactional customer enquiries to maximise organisational productivity, but customer service is a quintessentially human undertaking. Neither automation nor AI can ever fully substitute humans—capable of empathy, flexibility, and emotional intelligence—who can manage the nuanced interpretations that are nearly always required to generate the best customer service outcome.

Used properly, new tools like ChatGPT and the more advanced, AI-powered chatbots now coming into existence will make human agents more efficient, from improving contextual help in self-service channels, or improved decisioning in process automation, or indeed with predictive capabilities in more subjective areas like how people might respond to given situations or circumstances. In turn, this will make customer service a more productive and more pleasant experience for both agents and consumers alike.

The Limits of “More is Better”

Another potential pitfall to avoid is the notion that having more channels in customer service—including those that use AI-powered chatbots—is always better. In recent years, customers have changed how they want to engage with brands and organisations have responded by adding a variety of new channels. Where once the phone was the only way to get customer service, consumers can now choose from multiple channels including provider websites, social media, and in many cases, online chat. But these new customer service channels were not conceived to or made to work together as part of an overall offering, and customers have suffered the consequences of the lack of “connectivity” between them.

For example, a customer may begin a service enquiry on a brand’s website, and, failing to get a useful response, turn to the site’s “chat” option (which may be an agent but is more likely to be a ‘bot). Still unable to get a satisfactory response, the customer may then reach out via Twitter or Facebook, and possibly come up empty there as well.

Ultimately the customer reaches out by phone—meaning all those alternative channels failed in their purpose of steering customers away from this method of communication. In fact, data exists to show that most transactions that begin on channels other than the phone still end up as calls. So in the end, the proliferation of new customer service channels results in a more complex customer experience.

ChatGPT as Pathbreaker

ChatGPT is challenging the limits of chatbots, making them more responsive, reactive, and even proactive in the service of customers. Whereas earlier chatbots could only tap into a limited knowledge base and adhere to a rigid script to deliver limited responses, ChatGPT is expanding the scope of what chatbots can provide. In a sign of things to come, Microsoft recently made a significant investment in ChatGPT to boost its Bing search engine with ChatGPT’s advanced data processing capabilities.

ChatGPT has the potential to transform simple chatbots into sophisticated tools capable of learning and applying context—and in some cases even being able to understand emotion in a limited way. Natural Language Processing (NLP), whose capabilities have expanded by leaps and bounds, may eventually allow AI-powered technologies to guide live agents through real-time conversations in a way that was unimaginable just a short time ago.

AI as Transformer

The new AI technology will unlock the potential of data to deliver more accurate predictions and more contextual insights across a number of performance metrics. This new capability will improve self-service customer experiences and also provide agents with stronger support and more precise intelligence to solve complex customer service issues.

Customer service leaders should continue to steer the most transactional service requests to chatbots and leave more complex customer interactions to human agents. They should also focus on leveraging AI to provide the stronger support agents need to effectively manage those higher-stakes interactions. That support will come from AI applications specifically designed to streamline training, coaching, time and stress management, and other steps to boost agents’ skills and engagement.

All in the Service of Humans

Thoughtful placement and utilisation of this new AI technology is as important today as it was when chatbots were initially added to the customer service channel mix. Leveraging the full power of data will be key to acquiring the benefit potential of AI-based technology. Other essential steps must accompany or even precede the adoption of AI, such as moving to a Cloud-based data strategy.

And finally, regular calibration of AI models and feedback from chatbots will also be necessary to ensure that organisations reap the full rewards of AI in customer service: not as bloodless virtual agents, but as highly effective tools to elevate human agent performance and ensure consistently satisfying customer experiences.