The emergence of conversational UI
11 Apr 2016
If you’re looking to look at what’s likely to happen next in our use of mobiles, there are worse places to look than China.
While the market may have some differences – particularly in ecommerce, a lot of the features that have been seen in apps such as WeChat or Weibo have been starting to make their way into the type of tech we see in the west. In particular Chinese chat apps are engulfing more and more services – something we’ve seen for instance in the ability to hire an Uber from within Facebook Messenger.
Lying behind these innovations seen in the Chinese market is actually not a new idea at all. This is the rise of conversational UIs – user interfaces you converse with rather than click. In essence this is a new label for an old idea – chat bots as a concept have existed since the Turing Test was formulated – just a few years after computers were invented.
But whilst the ideal of conversing with a computer has always been there, often as a mere novelty, in recent years, the idea has seen renewed interest – particularly within mobile. There are three main reasons that lead to it coming to the surface…
- A narrow focus – These aren’t an attempt to make Skynet or Hal. They have their area of expertise and that’s what they’re sold as and optimised for. They’re not Cleverbot. They’re narrow AI. They do one thing well.
- Services behind the scene – If you want to order pizza through chat, it’s not going to work if the app has to then call the branch afterwards. Instead having well developed APIs that allow everything from taxis to share dealing to be ordered by computers is key to the growth of the approach.
- Big Data – Related to the growth in AI is the growth in big data sets and computing to analyse data. Want to understand what people are looking for? Build a big dataset of interactions and use this to optimise over time what you provide them. Ally to AI and it’s particularly fruitful.
So if this is going to be the thing in many a marketing plan for the next few years, what are the implications? To me one of the most interesting is what it does for branding.
To explain, over the past few years, anyone who’s had to formulate a social strategy has had to wrangle with the fact that the standard branding guidelines or tone of voice documentation only gets you so far you’re dealing with what to say on Twitter every day or chatting with customers on Facebook. Some of this is customer service sure, but it’s stretching beyond the remit of what brand guidelines would traditionally encompass.
Now consider Siri or Cortana (or their predecessors such as Ikea’s Anna). These agents of Microsoft, Apple and Ikea respectively are brand personalities – emissaries of them made (virtually) real. When Cortana banters with Clean Bandit in an ad, it is arguably the brand itself directly holding a (cringey) conversation.
This is now the challenge and opportunity that conversational UIs offer. Branding aside, one of the big opportunities, particularly for financial services is to provide added value to their customers by giving them access greater knowledge anytime, anyplace, anywhere. This is referred to as the robo advisor concept and it’s one banks such Barclays among others are embracing wholeheartedly. Want to ask your bank about invoice financing terms at 3 in the morning in your PJs? No problem.
But banks are also subject to compliance and regulation. It’s unclear yet how this meshes with conversational UIs. Is there going to be whole sections of knowledge or questions where users essentially run into a “I’m sorry Dave I can’t do that” moment like Hal in 2001? Moreover as users bump up against the edges of what the software can parse – remember, this is narrow AI designed for a specific purpose not a freewheeling conversation, we’re likely to see a lot of uncanny moments, unplanned results and downright comedy. Planning for the unexpected can go a long way – as Microsoft found to its costs when it recently when user interaction ended up making its chatbot turn racist in about three hours.
I expect that a lot of organisations aren’t going to think of these consequences in the rush to innovation. Chat can work well but expect to see the unprepared ending up with stranger, wilder unintended consequences, both technologically but more importantly with wider reaching impact for their brand.
Those using it successfully will think through what it means for your organisation as an emissary of it. This means working through any unintended consequences from unleashing it on the world - given it’s arguably one of the most personal embodiment of your brand you’ve ever created.
By Neil Major, Senior Strategist.