AI is the new language for talking with customers

Fears of a ‘techcession’, driven by inflation, conflict or simply coming off a pandemic-induced sugar high are easing. Indeed, earnings calls in February 2024 showed tech companies were buoyed by a return to spending on IT by businesses.

What specifically they are spending on is less clear, but the growth of artificial intelligence (AI) suggests businesses need to think about its potential impact on their customer experience (CX) and user experience (UX).

Ciarán Harris, co-founder and director of design agency and user experience specialists Each&Other said that the last few years have been interesting in relation to spending. First, businesses pulled back, concerned about the macroeconomic picture.

“It’s early days yet for 2024, but at the start of 2023 there was a lot of paring back. That seems to have picked up again toward the end of 2023: people are trying to understand how they can use AI to improve the customer experience. Certainly, we were asked a lot over the last year about how to leverage AI,” he said.

AI has a direct impact on customer experience, particularly when it is used to ensure that the customer journey is as smooth and frictionless as possible, thus leaving staff available to deal with complex and edge cases.

“One benefit is having a little bit smarter technology helps. It allows companies to take the reliance off some of the staff and use the staff for more complex tasks and also to be available 24/7,” he said.

Most of us, however, will have had the frustrating experience of being forced to perform simple tasks in the old-fashioned way.

“I was [recently] making a change in my travel insurance and I had to do it between nine and five, Monday to Friday. That’s bonkers. I was told to call a number and thought; ‘this is so antiquated’,” he said.

Naturally, then, one key area for businesses to investigate is customer service chatbots, a technology that is increasingly intelligent and can now be developed on platforms that don’t allow customer data to be used for training.

A lot of companies are wise to the fact that it’s customer experience that really does make or break how well their brands work.

The issue of data compliance is a real one, and the reason why many businesses have blocked access to public, web-based AIs and are casting a sceptical eye at the promises of the tech giants.

However, Harris said, there are methods for deploying AIs that can both be knowledgeable about a business and its products and, at the same time, not leak data.

“Mistral would be a good example. To train an LLM [artificial intelligence large language model] costs tens of millions of dollars, a lot of time and a lot of processing power. You start with a base model and add your data on top. That’s one way. The other is to take a model and give it intensive prompting and use that as the way to create the basis for knowledge about your product.

“Others are using the enterprise plans which guarantee, apparently, that your data is not being used for training. However, that really is one of the biggest concerns companies have.”

The coming change

In the slightly longer term, AI also has the potential to radically change how we interact with computers, which today means how we interact with a great many products and services.

Indeed, the whole graphical user interface (GUI) paradigm, which descended from work by the likes of Douglas Engelbart as well as Xerox Parc and was popularised by Apple, could potentially give way to something very different: conversation.

“‘Pictures on the glass’ has been our main modality since the 1970s. That is changing, but one of the things is we’re in a very fluid time. You can compare it to the first Macintosh or the GUIs Steve Jobs saw at Xerox Parc: we know this will change everything, but it’s hard to imagine how significant the change will be,” Harris said.

The key will be to get conversational interfaces to a place where it feels seamless and natural.

“We’ve been using OpenAI’s speech interface and in its ‘natural’ state it does have limitations, but you can train it,” he said.

In the meantime, Harris said, a lot of work is going into thinking about how to deliver on the promise of AI in CX.

“A lot of companies are wise to the fact that it’s customer experience that really does make or break how well their brands work, so there is a big push for people to implement AI – with the focus on the customer that can work well.

“We’re coming to the point where true personalisation is a possibility, especially for bigger companies. The challenge for SMEs is how they can keep up, and that’s where we see a lot of queries coming in: ‘How can we focus on our customers without the staffing investment of large corporations?’.”