Adding a human touch to artificial intelligence

The use of AI is becoming more and more common in everyday life, including in customer interactions with utilities providers. Yet on its own, AI will not be enough to provide the increasingly complex needs of consumers. Writing for Utility Week Rahul Arora, senior vice president and head of emerging business unit UK & Europe at data company EXL, argues that properly managed AI should augment human understanding.

Since the release of ChatGPT, AI has dominated headlines around the world. Where this time last year, the average person seemingly gave little thought to what powered their Amazon Alexa or their phone’s face ID function, suddenly AI is being discussed everywhere from dinner parties to the school playground. It’s no surprise then, that the topic has burst onto the boardroom agenda with similar force.

But it’s not a novel concept. Decision makers at the top table will know that AI has been quietly reshaping customer service for some time – from chatbots to personalised emails to voice recognition. This surge in interest – and simultaneous acceleration in development – is simply an invitation to give AI the prioritisation it deserves; to think seriously about whether it’s being implemented in the most effective ways, what more it can offer and how to ensure responsible use.

The utilities industry has always tended towards risk aversion. A good thing considering the services it provides are nothing short of essential for millions of households around the UK. But now the train is leaving the station and the choice to not act on AI is no longer an option. According to EY, 92% of energy & utilities companies are either already invested in AI or will do so in the next two years. The true differentiator will lie in how well AI adoption is carried out.

Above all, capitalising on the propulsive force of AI will come down to balancing the artificial with the real. Utilities customers will be delighted by the improved customer service AI brings – as modelled by the e-commerce industry – but no one wants to speak to a robot about their debt troubles or feel that no real person is keeping track of their situation.

This is a pivotal moment for utilities providers, and getting it right is vital. The companies who excel will be those who let AI do the work behind the scenes, but keep a human “in the loop”, to act at precisely the right moments.

The inflationary context: A human touch in debt management

The head of the International Energy Agency (IEA) recently warned UK households to brace for another surge in bills this winter. Clearly, as the energy crisis prevails, consumers’ pockets will remain stretched.

Debt management, issue resolution and treating customers fairly and responsibly puts huge pressure on customer contact functions, with inbound enquiries at record highs since the start of the energy crisis. Yet the traditional contact centre operating model relies on employing enough representatives to handle calls in an efficient manner – a model which longer works.

AI, analytics technologies and predictive modelling can help energy suppliers not just manage customer debt but – crucially – identify those at risk of falling into debt and work with them to prevent it from happening in the first place. And it can be trained to ensure that debt collection processes comply with regulations and laws, reducing the risk of non-compliance and potential penalties, as well broadly upkeeping fair, legal, and ethical conduct.

But debt management is a sensitive topic. For some customers, the idea of their energy provider leaving it to technology to trawl through a myriad of data sources to make a decision on whether they may struggle to pay their bill, may be distinctly uncomfortable. And they almost certainly do not want to be conversing with a chatbot in detail about their financial challenges.

Energy providers must engage with the experts – whether external or hired – to identify where AI can make a difference in protecting their customers from debt and where necessary helping them manage it, whilst understanding at what point in the customer journey to involve a human customer service operative.

Properly managed AI should augment human understanding. On its own it can do many things, but it can’t provide human empathy and understanding. The goal of any AI implementation shouldn’t be to replace workers, but to let employees focus their time on tasks that only a person can perform well.

A targeted approach: Knowing your customer

Looking beyond the sensitive topic of debt to all aspects of customer service, not every customer is the same. Some will be more receptive to interacting directly with AI than others. Digital natives may feel more comfortable using self-service chatbots. Users facing complex issues may prefer speaking with a human representative right away.

Fortunately, AI can assist in informing decisions about its own deployment. It enables the ‘joining up’ of information about customers; both structured data such as information held on CRM databases, and unstructured data from existing sources such as recorded telephone calls, or emails. Analytics technologies then help human agents make real time, data driven decisions and importantly, give them a complete customer view. This helps serve the customer better in a hundred different ways, but among them AI can tell which customers will require human attention more often, or earlier than others. The utilities providers that get this balancing act right will be those which build channel strategies around customer preferences.

The opportunity to deliver best-in-class customer experience has never been greater, but with more customers struggling to pay their energy bills than ever before and the regulator keeping a watchful eye over the entire UK energy sector, providers must move not only with care but with speed.