Ask the leadership of any reasonably-sized company what technology they’re looking to implement and they’ll almost invariably mention artificial intelligence (AI).
In theory, that’s great, because AI has the potential to fundamentally change the way a businesses operates and creates a great Customer Experience. The longer the business uses an AI application, the better the experience should get. Given enough time, the system can collect enough data on each individual customer to provide meaningful, hyper-personalised experiences.
Implemented badly, however, AI can be a total disaster. Rather than feeling like the business they’re dealing with cares about them, they’re left with the impression that customer service has been handed over to a bunch of dimwitted machines.
Let’s talk about chatbots
The easiest way to illustrate how varied the AI experience can be is to look at chatbots. They’re the kind of front-facing AI that more companies are using and which an increasingly large body of customers are familiar with. Trouble is, most companies are terrible at implementing chatbots.
Apart from a few forward-thinking exceptions, companies tend to put a chatbot on their website in the hope that that it will learn from each interaction it has with a customer and that its answers will become more nuanced over time. They also operate in the belief that customers will tell the chatbot when it’s wrong, helping to train it further (hands up if you’ve ever done this willingly).
That would be great…if the chatbot was actually equipped to do so. However, for the most part, chatbots are simply going through the company’s existing knowledge bases and serving you with a document (or, in the worst cases, multiple documents) to try and help. It’s essentially a slightly smarter form of search.
As anyone who’s tried to use the search function on a corporate website will tell you – that’s not particularly helpful, especially when you’ve got a specific query. Let’s say that I want to know if I can insure my sunglasses. I don’t want to have to scour through insurance agency documents to try and figure out the answer. I just want the answer.
Contextual, hyper-personalised, relevant
As long as chatbots rely on a flawed architecture that depends on the existence of relevant documents containing the needed information, they won’t be able to provide that answer.
If you’re going to use AI to improve CX, you need to take a different approach. If you want to operate in the digital era and want to drive logic through data then you need to start it in data. That means looking beyond your existing documentation and CX architecture and integrating insight into customer behaviour across digital and offline channels.
This approach will, ultimately, allow you to offer customer support that is hyper-personalised, relevant, and compliant.
A chatbot built on this kind of framework understands what you’re asking and can answer specific questions according to what you actually need. While that’s just one small part of CX, anyone who’s cursed a company for failing to provide useful information, will know how important it is.
The aim of AI
That said, this approach shouldn’t be limited to chatbots. Consistency – in style, tone, and content – is one of the most important factors in successful CX.
It’s therefore imperative that any organisation turning to AI to improve CX apply a data-first architecture across every customer-facing channel. So, whether I make a query using a chatbot, the search function on a website, or a call centre, I should get the same – relevant – answer.
However, if this is going to happen, businesses need to stop trying to bolt AI onto their existing architectures and take an approach that allows it to reach its full potential.