As the world moves toward greater automation, chatbots seem like a great way to cut costs on customer service personnel. But before you leave the bots in charge, consider the potential ways chatbots may turn on you.
1. Not answering the question at hand
Anyone who’s dealt with chatbots knows the frustration they can induce. Sure, bots that deliver news stories – like CNN’s bot – are fine, but bots meant to bridge self-service with customer service don’t always get the job done.
With limited commands, it can be difficult – if not impossible – to get the information you’re seeking. In such moments it feels like you need a secret password – and no one’s sharing.
Perhaps consumers who use chatbots regularly are better able to navigate these scenarios, but if new users can’t easily figure things out, there’s not much incentive for them to stick around.
2. Learning from bad human behavior
To humanize the chatbot experience, bots like Microsoft’s Tay, use machine learning so the bot becomes “smarter” the more people use it.
Machine learning is a beautiful thing – when it works. What Tay showed us is an artificial intelligence (AI) glitch that needs to be worked out. Your audience may not care that mischievous social users sabotaged your bot and taught it to make racist comments – they’re just going to care about the impression of your brand they experience in the moment.
3. Not alerting a real human soon enough to salvage a problem situation
When things go wrong, chatbots are supposed to be intelligent enough to alert a human to take over – but that doesn’t always happen quickly enough.
Though the idea of chatbots is to give consumers another avenue to manage their needs, it can be a lot of work to get a bot to understand what those are. And when you’re frustrated or angry about a bad experience, this just intensifies the feelings of dissatisfaction.
Chatbots are useless if they can’t alert your brand to the need for human intervention when things start to crumble.
4. Giving consumers the impression you don’t care
Much like automated phone services – something consumers have grumbled about since they came into play – chatbots in the customer service realm feel like just another hoop to jump through. They tell consumers you don’t have time for them, so you’ve offered up a work-around to avoid dealing with them.
What these “betrayals” have in common is a lack of something consumers have increasingly demanded in the social age: personalization. While chatbots that deliver news, sports scores, flight updates, etc. can be personalized to give consumers what they specifically want, customer service chatbots represent a regression for brand-consumer relations.
Though they’re meant to serve as a triage coordinator, they too often fail to get consumers to the next level – unless they take the next step themselves by calling, emailing, or reaching out via social channels.
All this begs the question: Do consumers even want to deal with chatbots?
Taking consumers’ temperatures
If that question hasn’t occurred to you before, that’s a problem. With all things, brands must seek consumer opinions – and analyze that feedback – or suffer the consequences. The simple fact a technology exists does not automatically mean your audience is ready to embrace it.
If your social listening and sentiment analysis doesn’t point to a love for chatbots, you don’t need to use them. If your audience has an interest in hearing about new products or tips and tricks via a chatbot – but not using bots when they have a problem – that’s worth paying attention to.
Another tactic is to use your social monitoring software to gather a little competitive intelligence. What chatbots are popular with consumers, and why? Be sure yours lives up to the standards your audience expects.
AI assistants may someday be able to reliably stand in for humans, but we’re not quite there yet. In the meantime, let consumers guide the way you introduce new technology, and don’t let your chatbot sabotage your brand health by thinking it’s ready to fly solo.
Want to know more? Get in touch for a demo of our social analytics tools in action.
Image from Adrian Brady