Cringe-inducing chatbots are the perfect example of why technology isn’t always the answer

Chatbots aren’t human and when they pretend to be, we hate them in a way that only humans can, writes Hamish Cargill.

Every so often, something launches into the world so quickly and completely that you can’t remember if it was always there. One Direction is a great example. So is Pokemon Go. More recently, fidget spinners. One minute nothing, then bang – the whole world is going nuts.

That’s how I feel about chatbots. Not so long ago, they were just a toy for tech nerds – distant, undefined, obscure. Like the Apple Watch, they were a technology in need of a reason to exist, and I’d argue they are yet to find that sense of purpose.

With the help of a few pioneering brands, in the first half of 2017, the not-so-humble chatbot has embarked on a steady march to becoming mainstream. It’s hard to escape the howls of excitement telling me chatbots are going to change my life, your life and most probably the world. But are they? Are they really? And what the heck does all this have to do with me, a guy that looks after brand voice for a living?

Not a chatbot (AKA Hamish Cargill)

Since chatbots are about as visually appealing as mouldy cheese, the strength lies in what they say and, even more importantly, how they say it. That brings it into my orbit and raises some important questions. If chatbots are now doing all the talking, what conversation are they having, whose voice are they doing it in and does anyone even want to talk to them?

The human/machine dilemma

Chatbots aren’t human and when they pretend to be, we hate them in a way that only humans can. That’s something we’ve learnt already – when real people think they’ve been talking to a person but find out they’ve been talking to a robot, they understandably feel cheated, tricked and creeped-out. A good chatbot should be like a human, but it can’t pretend to actually be one.

Since chatbots can’t demonstrate the range of human emotions of even the most poorly trained call centre worker, they’re best kept well away from situations that require empathy. This includes anything from emergency hotlines, through to insurance claims, overdue bills, debt collection (here’s looking at you, Centrelink). These aren’t areas that brands have traditionally excelled in, and a bot’s certainly not going to improve the situation.

Don’t let those smiles fool you

But not every interaction between people and brands is an emotional one. Most are just plain transactional. Chatbots are much more suited to this line of work. If there’s no added value that a human could bring, then there’s no need to invite one to the party. Information sharing, processes, online checkouts, regular bill processing – none of us expect empathy from these situations.

Dominos have nailed it with Dom the Pizza Bot. Given the low-risk context of this interaction, it’s the perfect environment to test this technology. He wins early by acknowledging bot status and brings a consistently cheeky and playful tone to the limited process of ordering pizza (only with a topping you’ve already bought before) through Facebook Messenger. Dom’s job is a simple one – he’s not expected to resolve complex problems, and he’s been given the ability to openly say when he’s overwhelmed. That makes him a success.

Plus, Dom the Pizza Bot sends GIFs…

Ebay has tried something similar with their ShopBot. It helps you find styles and products based on the criteria you select. This is a more complex area for a bot to work in, and the repetitive nature of the experience makes it feel stilted and slow. Tonally it’s fine, acknowledging flaws and constantly asking for feedback. But in truth, it’s more difficult to use than the traditional way of searching Ebay. It’s a bot for bot’s sake and just another reminder that technology isn’t always the answer.

Bot language

These examples show the strengths and weaknesses of bots. Critically, they demonstrate the responsibility that language – and more importantly tone of voice – are carrying as the technology struggles to keep up with the hype.

That’s where this issue of ‘faking it’ becomes so important. It’s not enough to make a bot that’s chatty and cheerful and charming if every other interaction your brand is having is as dry as day-old toast. That’s fake, and people smell fake like they smell a fart.

In an ideal world, your brand would have such a strong personality and voice that people just ‘know’ you. That way when you add a chatbot to the mix, it will be written in your brand voice and can’t be taken out of context. People ‘get’ you and know what you’re saying and why you’re saying it.

It would be easy to think of chatbots as new technology that deserves special treatment. In reality, this is just another means to communicate with an audience that demands more from the brands they interact with than ever before. As always when brands communicate, a successful chatbot will demonstrate distinctiveness, consistency, and an ability to connect with that audience.

When they can do all that, chatbots might finally be worth talking to.

Hamish Cargill is the director of brand language at XXVI.


Get the latest media and marketing industry news (and views) direct to your inbox.

Sign up to the free Mumbrella newsletter now.



Sign up to our free daily update to get the latest in media and marketing.