This piece by our Managing Partner Sarah Cantillon originally appeared in The Drum.
Chatbots were predicted to be prolific. And for a while they appeared to be; circa 2016, chatbots became the most prominent tool on the great marketing bandwagon. Marketers went to SXSW, drank the Kool-Aid and decided they need to build a bot or be left behind. Businesses of all shapes and sizes started making them in their droves, best chatbot categories started springing up in industry awards.
Fast forward to today and, despite many bots languishing pointlessly or worse, annoying customers, clients are still pushing digital agencies (us included) to make more. But there’s often little thought going into how bots might affect brand experiences, or most importantly, why companies actually need them.
Why a chatbot?
Before building a chatbot, first ask yourself why. It sounds obvious, but it’s surprising how many businesses get carried away with the idea of a bot before considering the practicalities. It’s not just a question for a developer and UX designer – planning is crucial too.
Test thoroughly beforehand to see if a bot is the right solution. From our own recent primary research for a client on a specific project, we found that while people liked interacting with prototype chatbots as part of an experiment on their website, due mostly to the fact they’re a novel experience, they’d be unlikely to use them in real life. They were considered an uncomfortable fit for the scenario. We told the brand that a bot wasn’t the way to go in this case and offered alternative, more suitable, approaches.
Bots stripped back
So should all brands forget bots? Not necessarily – there is a role for them, but it’s often less exciting (and therefore less expensive) than most marketers believe. Bots usually need to be stripped back.
They are most useful when dealing with finite, limited, closed questions rather than open questions. Much like automatic supermarket checkouts, bots need to forget the niceties and just give consumers what they want conveniently and quickly. They don’t need to try be ‘real’ – at the moment, mass, mainstream AI hasn’t evolved to the level where your average online retailer can add a bot to their site that can converse with a customer to the same level of sophistication as a human. It grinds a customer’s gears when they want to speak to a real person and a bot is wasting their time with false friendly language – and even worse if a basic bot is purporting to be a human. Being open with the fact your bot is just that, will manage customers’ expectations.
Where they are hosted can make a difference. They can feel more ‘part of the furniture’ on social platforms than on websites. As of June 2018, there were more than 300,000 active chatbots on Facebook Messenger. They’re easy to build and customers on Messenger are expecting them – they’re a familiar part of the platform for many users and people actively turn to them for customer services purposes.
The act of seeking out a bot is also important – part of the reason why people are often uncomfortable with them is when they feel interaction is forced by insistent pop ups on websites. Bots should sit in the background, ready to be used if needed, staying quiet if not.
Not all bots are bad. And as with most new technologies, people will get used to them and become more willing to interact with them in time and as the AI improves (unless they’ve been usurped by something better), so it’s still worth becoming familiar with their capabilities and limitations.
But they are not always a go-to solution for easy online customer interaction. They can add an unnecessary layer of complexity when brands should be aiming for simpler experiences, not more difficult ones. Businesses need to remember that the best customer experiences are based not on what you add, such as a bot, but what the consumer takes away. If they’re left frustrated by your website, it’s time to re-think your digital approach.