Even as we grow more and more dependent on digital technologies, people still like to interact with each other. However, in our increasingly digital lives, we often need access to information more quickly than before. And visiting a person or making a phone call for help can’t always deliver what we need as fast as we need it.

That creates a challenge. How do we exchange information with a client or partner quickly when we’re inundated with data and requests? How do we know what they are looking for specifically? The exact data they need may be the proverbial needle in our data haystack.

The proliferation of technology has given rise to the challenges of more contact avenues, increased data volumes and expectations of fast service. However, it also creates opportunities to differentiate your business from that of your competitors.

Artificial intelligence (AI) and machine learning can help. For example, if a customer is looking for travel insurance, they might browse your website in search of a product that meets their travel needs. Understanding what they have browsed on the website helps you anticipate what they may need.

When the customer contacts a sales associate over the phone or via online chat, the associate can already have preloaded the information needed to help the customer. Rather than replacing the person on the phone or at the other end of the chat, machine learning and artificial intelligence can complement the associate’s service skills, allowing them to deal with everyday queries more efficiently and freeing them up for more challenging tasks.

The next step is to have the technology assume the role of the sales associate. Machine learning and AI can be used to drive chatbots. However, these need to be carefully used as, despite being efficient and effective, they often leave people feeling like they have been shifted from a human interaction they see as valuable to an electronic one that is less satisfying. 

But that is changing. While a new generation of customers actually prefers non-human interactions, companies are working hard to imbue chat systems with more humanlike traits. This means not just having chatbots speak with a more human voice, but also better understanding the nuances of human language. For example, an Australian utility recently started training its chatbots to understand subtle cues and subtext when people were understating their problems. Others are using sentiment analysis to understand sarcasm and other tricks of human speech.

A number of technologies and tools are emerging to help people better adapt and accept AI-based systems in their interactions. Some companies are using human actors who are technologically augmented so the chatbots are given a human face.

Instead of computer-created voices, such as the one Stephen Hawking had in the computer he used to communicate, more lifelike voices have been developed, and many chatbots are given familiar, uncomplicated names to help suspend the disbelief that you’re not talking to a real person. Just look at how many people can identify Siri or Alexa by their voices. We are already starting to anthropomorphise these computer-based tools. 

So where can this technology be used? In short, almost any field in which a customer asks a question is a candidate for the application of a digital human powered by AI and machine learning.

In finance and insurance, a digital human can answer routine enquiries and pass them on to a human operator when the technology can’t resolve the question. In retail, a chatbot can help someone choose the right clothes. For example, customers could provide the system with their body measurements and the occasion they’re dressing for, and a chatbot could then guide them through an online store to help put together the perfect ensemble.

When a customer books their next holiday, they could tell a chatbot what sort of holiday they want, their budget and preferred travel dates, and the chatbot could suggest some potential locations, refining the choices based on the “discussion”.

The Turing Test was developed almost 70 years ago as a test of a machine’s ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. Passing the Turing Test is not the goal of today’s digital humans. The goal is to make the service outcome of interacting with a bot no different than, and hopefully better than, dealing with a human. 

We are on a journey towards that outcome.