Is AI overhyped? How Automation is the Real Customer Service Hero

If you haven’t noticed, ‘AI’ is everywhere right now. Since 2016, Google Trends has seen a steady rise in searches for ‘artificial intelligence’; searches for ‘machine learning’ are much the same. AI is everywhere — it’s been called the future of customer service (CS) and a threat to the job safety of human operators. But is it overhyped?

Well, Gartner’s 2017 Hype Cycle firmly places ‘deep learning’, ‘machine learning’, and ‘virtual assistants’ within the peak of inflated expectations. And this is, in part, because of the misinformation spread about what ‘AI’ really is and what it can do. AI is often misconstrued with customer service automation, which is simply the automating of various processes within your customer care. Real artificial intelligences are more complex and they’re not ready to take over the customer care journey yet — they probably won’t be for some time.

AI and Automation: The Difference

According to Gartner, an AI must improve from experience, teach itself when given new data, and produce unexpected results. However, most robots that we see in CS right now are programmed by developers and only act based on the code they’ve been given.

So, what does a real AI look like? There are plenty of companies trying to produce AIs at the moment. Facebook recently shut down two AI robots when they started creating their own language to communicate. Facebook were hoping to create bots that could converse naturally with humans. But, instead, when they asked the bots to talk to one another, they quickly started making up their own language. And this is just one of the examples why the technology behind AI just isn’t ready.

Even when the bots are successful in their assigned tasks, it’s still clear that AI is years away; Google’s deep learning robot recently taught itself to walk. Whilst amusing and — arguably — impressive, it isn’t the ‘AI’ that everyone has been talking about for customer service. In fact, at the moment, there aren’t even any real examples of AI in customer service; even virtual assistants like Amazon’s Alexa are not ‘true’ AIs (though it could be argued that they have augmented intelligence).

Automation, on the other hand, is proving incredibly useful for businesses. Through the use of workflows, automation can prioritise interactions (including prioritising important or difficult queries), route conversations to appropriate operators, send follow-up surveys, and even reduce spam. Automation isn’t about to take over the entire customer journey but that doesn’t mean it isn’t helpful.

But what about chatbots? Aren’t they AIs? Well, not really, no. Chatbots can’t think for themselves without receiving programming. They can have some form of augmented intelligence but most businesses who are utilising chatbots simply don’t need them to be that smart. At their core, chatbots are just automation software with a natural language understanding engine, meaning they can decipher what a customer is saying and formulate a reply that’s been programmed into them.

But — considering how far away AI is — that’s arguablly more valuable for a company that deals with a high number of contacts than an AI currently would be. Our own research found that up to 78% of retail customer queries could be answered by chatbots. Most queries were FAQs (like ‘where is more order?’) or were asking for a return or refund. Chatbots can handle this kind of interaction without impeding the customer experience.

Will automation steal your job?

Since chatbots can do their job so well, is it likely that they will replace human operators? In the long run, maybe. But that would require bots with AI and — as I said before — we’re still quite far away from that reality. At the moment, automation works best alongside human operators, not without them.

Although a chatbot can handle 78% of a business’ contacts, that still leaves 22% of the queries. Those interactions are more complex and almost always require a human operator to resolve them. Chatbots simply aren’t smart enough to answer every question.

They’re also not emotional enough. Chatbots can’t feel emotions — nor can they detect them. Excellent customer service, at its heart, is still all about empathy. If a customer is upset or angry, chatbots can’t respond to this and they can’t reassure the customer — just another reason to keep a human team of operators.

Automation is designed to increase efficiency. Whilst that may sound like a veiled threat to an agent whose job could be on the line, it’s actually an operator’s dream. Using automation software like Gnatta means that operators can have all the information they need in one, easy-to-use interface. Chatbots can collect the data for them and the software can store all historical data from previous interactions; all interactions, no matter the channel, can be collected in one place so it’s simple and fast to click between one conversation and another; and Content in Gnatta can provide the operators with suggested, customisable replies whenever its appropriate. All this means the operator can work more efficiently and have more time to focus on providing an amazing experience to each customer. That not only improves the brand, it actually also adds value to the very job role of the operator.

How is automation helping companies now?

So, whilst AI isn’t the answer right now, automation is. And there’s tangible evidence that it’s working. Gnatta recently conducted a study into the benefits our clients have seen since using Gnatta’s automation software. We saved one client over 3,000 hours of labour per 100,000 interactions and increased their interactions per hour from 12 to 20. That’s a cost saving of £26,000 a year. And that was just by increasing productivity and efficiency with automation.

For another client, we focused on reducing their contacts by implementing profile aggregation and automation, assigning all inbound contacts to a unique user profile. This saved the client £3.2 million a year!

Why do we say ‘AI’?

If true AI technology is still years away, why do companies say ‘AI’? Even we at Gnatta are guilty of it. The reality is that the concept of AI is a complex one. There’s a difference between augmented intelligence and artificial intelligence (but they can both be referred to as ‘AI’); some technology — like machine learning — is ready now but the next steps — like deep learning — are still in development; even the definition of a chatbot is debateable. So, in reality, the term ‘AI’ ties everything up with a bow and is a neat, little term to refer to technology that’s more developed than a simple bot and less developed than HAL 9000.

So, is AI and automation the bogeyman it’s been cast as? Are they the heraldic saviour of customer service? Are they here to uproot the job market for customer service operators? No, no, and no. But, can automation and chatbots offer impressive efficiency gains, save a company money, and make the job of the operator easier and more effective all in one? Yes! You just have to work with the right company for the job. If you’re interested in automation or chatbots, get in touch with us today. Alternatively, check out our automation whitepaper here.