Consumer interests have changed, and they will keep changing.
Social media switched the balance of power away from company-controlled communication and into the consumer camp — not just in-the-moment, but precisely at their moment, on their devices, on the channels they choose.
But with the advent of messaging and the recent injection of A.I. into messaging platforms in the form of chatbots, it’s more than just communication that is shifting — it’s the entire way people and businesses interact. While messaging platforms are the future of consumer-brand engagement, A.I. technology is in its infancy as it relates to truly engaging with humans. There is a long way to go before people are having open-ended conversations with chatbots that don’t end in disappointment.
Just message me
Before I explain my point, here’s some background on messaging apps. Messaging platforms are clearly on the rise, growing faster than social media. The four biggest messaging apps are now bigger (in terms of monthly active users) than the four biggest social networks.
Yet, just as communication has made a huge shift away from older digital channels and into messaging, we’re about to see a major shift away from traditional mobile applications and into interactive functionality delivered seamlessly to customers inside messaging applications, often powered by A.I. and bots, with customer service fitting alongside.
WeChat led the way for this to some extent, providing a single ecosystem with payments, identity, and messaging where companies can deliver full app-like functionality within the messaging application.
Facebook Messenger hasn’t gone quite that far (yet), but it does allow interactive functionality to be delivered inside the messaging thread. It is enabling, for example, the ability to flick through a carousel of options, or to have check-in notifications and boarding passes delivered. Some of these bots are using basic natural language processing (NLP) on top of the interactive features, with the option to hand off to a human agent inside the same conversation if a consumer wants to ask anything the bots can’t answer.
Is this really A.I. at all?
Around the same time that Facebook launched their interactive “bot” functionality, a host of other major companies, including Microsoft and Google, released various new platforms that make it easier for developers to employ artificial intelligence and natural language processing in their own bots. With all this news came a ton of hype. For a couple of weeks, bots were the hottest thing in Silicon Valley.
But messenger bots today aren’t really A.I. They’re just a different UI accessed inside a messaging thread. It’s not text as UI — it’s UI as UI, but delivered inside a messaging thread with some conversational elements for text entry.
That doesn’t mean they can’t be useful — it’s easier and faster to order flowers using the 1-800 Flowers bot than it is to download a whole separate application that will then be immediately relegated to your third iPhone screen. For the same reason, I’d much rather have my KLM boarding pass delivered to me over Messenger. As this kind of functionality deepens, we’ll see more and more businesses opt to build messaging “bots” (or applets) versus traditional applications. With iOS and Android both improving their own native messaging and notification windows, this functionality could even be delivered straight to a phone without even a messaging platform in between.
Add in A.I. that combines personal data about a customer with situational intelligence from the phone, and the functionality you need could be delivered seamlessly just when you need it.
No smart chatbots (yet)
In the past few years, innovation — especially in hardware — has made large-scale deep learning widely accessible. We now have a thousand flowers blooming across startups and established companies, putting machine learning to work to drive innovation in every area imaginable.
Combining this with the ability to inject interactive functionality into every mobile engagement with your customer has the potential to make interactions with your customers faster, smarter, and more efficient. But this will come through the application of A.I. in narrow use cases — predicting certain types of issues, understanding customer behavior better, and potentially enabling customer service agents to do their jobs much faster and easier.
But when it comes to real, meaningful customer service conversations, the fact is that AI is just not good enough yet. If you try to use it for general open-ended conversation, it will quickly cause frustration and annoyance. Think of the Tay bot or consumers who turned to social because they were fed up with interactive voice response (IVR) and computers saying no. Replicating the same customer experience in messaging will guarantee that they’ll be back at it, complaining on social media fiercely — and publicly.