fakeitmakeit

It’s obvious that messaging as a new UI is becoming one of the new mega trends this year. Two weeks ago Apple announced with iOS 10 that it’s opening up iMessages to outside developers, while Facebook’s recent F8 developer conference kicked off both days with keynotes on bots and messaging. Microsoft, too, has opened up Skype messaging to third party developers through Cortana.

With all these big American aircraft carriers all moving to get messaging and bot infrastructure on their platforms, one thing is clear: there is now enough widespread support to give messaging its proper place as a ubiquitous, smart, and natural UI for harnessing technology and getting things done. So why are the big platforms suddenly so interested in ‘opening up’ their messaging infrastructure?

This inspiration seems to be coming straight from WeChat. Opening the app in China you can find every possible B2C businesses like takeout, banking, and taxis accessible in the app, just like any other contact. And you’ll see it’s as easy as a few taps inside WeChat to order & pay for your pizza, or arrange your Red Envelope (indeed a very big deal for WeChat Pay). Microsoft even launched a chatbot on the platform named Xiaoice, which counts over 20 million followers.

So instead of an advertising business like that of Facebook or any other traditional online media outlet, WeChat is making a fortune as a facilitator of _transactions_. Allegedly, the run-rate payment volume processed by WeChat is twice that of Paypal. Staggering.

There is good reason why you could say the WeChat ecosystem is what Facebook’s Messenger or Microsoft are hoping to build. But remembering WeChat’s legacy, it launched on a mobile-first audience in an environment where there was no other real app store. Like Gustav Söderström said so brilliantly, WeChat built the browser Marc Andreessen would have built if he started Netscape today: social, personal, and with seamlessly integrated payments. But in the western world of 2016, the incremental value of bringing transactions onto the Messaging platform is much smaller considering the alternatives.

The big reason why is that App Stores in the west provide already provide the infrastructure WeChat built for China. Although App Stores don’t count the social angle, their credit cards on file already cater very well to e-merchants, take-out restaurants, airlines or games. This is why the leading messaging platforms have a much tougher nut to crack than their Asian counterparts and are IMHO very unlikely to become the same transactional rocket ships. Best case, Facebook Messenger will move to capture a large share of the support tickets that otherwise pop up on company pages. But transactions – I think that’s a far cry away.

I expect over here this field will be much bigger than just a few giant tech platforms launching bot frameworks. Taking a look at Europe, a few on our radar include Comtravo – a business travel app that works conversationally. Coming to Stockholm to visit us? Just use your messaging platform of preference (like Skype or text) and get easy options for hotels and flights back in the thread. Instant booking just by sending another message. German colleagues Voya does much of the same but in its own proprietary app channel.

Two other notable apps with conversations beautifully integrated as a key part of the service (although neither have a ‘conversational’ core) are depop, where buyers and sellers of vintage fashion, apparel and other stuff message each other in-app around an item or transaction; and Virta Health, started by ex-Trulia COO Sami Inkinen, which partly uses conversational UX to deliver their individualized therapies to restore metabolic health in chronic disease patients. And illustrating some of the hype, we see tech companies in search of additional growth or cred acquiring ‘even newer tech’ from this particular scene.

These are just a few examples (and for disclosure: some of them are already in the Creandum portfolio) but let’s have a look at this map visualizing the different approaches to Messaging & Conversational UI:

graph

Of course, that top-right corner where you’re solving a vast range of use-cases with AI is what many dream of. And for good reason – an artificial intelligence almost by definition should make sense of and execute a user’s most complex demands with perfect accuracy. More than a few start-ups have been brought to the world with this ambition. However, given the difficulty of automating everything for everyone, I think this is an incredibly risky approach and one that I feel won’t be appropriately rewarded in a long time.

How I imagine this will work is that entrepreneurs will launch startups to target their own specific niche, which will require lots of handholding and in the beginning, basically faking an all-encompassing AI behind the scenes. This places these startups squarely in the bottom left corner. But as they grow, they will be uniquely well positioned to know what parts of their use-case and messaging to automate & how to do it, over time reducing need for human input and allowing these companies to truly scale. With this tech layer and market lockdown in place, I expect these startups to then diversify to reach a wider and wider audience.

This is the long game. It’s incredibly tough to create a messaging AI that does all things for all people, but the startups that get their niches right will learn the specifics of their interface with users to automate it well, and thereby have the right tools to take on a larger and larger slice of this industry.

I’m not arguing with the big players that AI and bots will be important. But in the medium term, app-assisted personalized interaction is actually the big deal. Learn about how to best design your interface to the user. Then figure out what parts of it lend itself well to conversations. Study those interactions. And then, armed with unique knowledge about your use cases and all the excellent NLP and Semantic Search tools at hand, set out to automate the parts you can.