Sunday, April 17, 2016

Messenger Bots: What Are They, How do They Work, and Why Should You Care?

By Andrew Hutchinson

Is your business talking about Messenger bots yet? They will be soon – according to a report from TechCrunch, Facebook is set to launch a Messenger Bot store at their F8 conference this week (UPDATE: Facebook has now announced the Messenger Bot Platform, read about it here). Why is that such a big deal? Because Messenger bots have the potential to change the way a great many brands do business –
sure, we’ve had automated messaging products before, but Messenger bots are different, more advanced, more personalized. And they’re coming into the market at just the right time to change the game.

WHAT IS A MESSENGER BOT?

Essentially, a Messenger bot is an AI-fuelled system that can respond to customer queries sent to your business via direct message. This functionality is not exclusive to Facebook Messenger – messaging platform Kik recently released their own bot store along the same lines – but Messenger has much higher reach (275m for Kik vs 900m for Messenger), and is thus, more likely to be of immediate relevance to your business.
Kik actually made a helpful video to explain bots as part of their bot store launch.



Bots, as noted in the video, are like people, but they’re not actually people, they’re computers. We’ve had bots in our phones for years in the form of applications like Siri, but the new generation of bots is going to be far more advanced and – most importantly – individually customizable to your business.

WHY ARE MESSENGER BOTS SO IMPORTANT?

Facebook’s Messenger bots will be more advanced that your common bot system because they’ll be able to utilize Facebook’s AI learnings to develop more intelligent and more intuitive systems. According to TechCrunch, Facebook’s Messenger bot store will help businesses build automated response systems for fielding messages from potential customers – but rather than putting the onus on business owners to learn how to construct such complex, code heavy applications, Facebook will refer them to a listing of approved providers who can assist, and who have a solid working understanding of how Facebook’s bot systems work.   
Much of these learnings will come from systems that Facebook already has in place – in this sense, one of the best places to look for an example of what Messenger bots could look like is Facebook’s AI help system ‘M’.
Launched back in August (for users in the Bay Area), M is Facebook’s ambitious play to make Messenger a more central part of everything you do. As explained by Facebook Messenger lead David Marcus:
Unlike other AI-based services in the market, M can actually complete tasks on your behalf. It can purchase items, get gifts delivered to your loved ones, book restaurants, travel arrangements, appointments and way more.”
M is basically your own personal assistant, able to perform a wide range of tasks for you – and the genius of M is that with every action it undertakes, its learning. Right now, M is powered by AI, but that artificial intelligence engine is overseen by a team of ‘trainers’, customer service experts who can step in to ensure M is performing each task being requested of it. The idea here is that, over time, and with each human intervention, M is learning where it went wrong, learning the subtle nuances of the queries being processed, learning, essentially, how to best respond to queries and work without people being involved at all. And it’s getting better – in a post about his experiences with M, BuzzFeed’s Alex Kantrowitz noted that it was often difficult to tell whether he was dealing with a person or a bot.
That kind of ambiguity is exactly what Facebook’s aiming for – if they can build an AI system that accurately mimics the responses of a real person, and completes all tasks as required, that could provide a huge boost for a wide range of businesses.
And here’s where the true aim of M may actually lie. M, in itself, is a difficult concept to scale – having an AI system with human trainers means you need a lot of people, especially if you were to be catering for all of Messenger’s 900 million active users across the world. Facebook has said that they envision hiring thousands of trainers, a huge team of people helping to fuel that AI engine and build M’s intelligence. But what if, through the millions of potential queries being processed by M, they were actually building intelligent bot systems that could be customized to various industries? 
As in the example shots above, what if Facebook could offer you a bot that was already able to handle the majority of customer queries related to your business with a high success rate?
Say you’re a pizza store owner and you go to the bot store and you see that there’s a bot on offer that’s already handling 75% of customer queries to similar pizza stores without human intervention? That’d be a potential game changer for your business – and when you combine that potential with Facebook’s new Messenger connection tools, like Messenger Codes and shortlinks, you can start to see how Messenger bots could change the game.
Add to that the fact that people are already using Messenger – the platform recently announced (and as noted earlier) that they’ve hit 900 million monthly active users. People are spending more time on their smartphones every year, as per this chart from Mark Meeker’s 2015 Internet Trends report.

Of that time, 85% is spent in apps – yet only five apps, on average, see any regular use. Combine that with recent findings from comScore which show that two-thirds of U.S. smartphone users downloaded zero apps in the last month and you can definitely see the app trend beginning to slow - while at the same time, 1.4 billion people used a chat app last year, according to eMarketer, a number projected to grow to 2 billion by 2018.
The data all points to people spending more time in chat apps, communicating more via direct message.
Given those trends, being present on the platforms where people are most active is where brands, logically, want to be.
Interacting with a brand via message is potentially faster, easier and more customizable than other platforms. And what’s more, for the brands themselves, if they can use bots to automate the process, it’ll also prove significantly cheaper.
This is why Messenger bots are so important, and how Messenger bots will change the game.

SO HOW DOES FACEBOOK MAKE MONEY OUT OF BOTS?

This is the golden question right now. No one’s 100% sure how Facebook plans to make money out of bots. Of course, there’d be a charge for creating a customized bot, there’d be a cost for accessing Facebook’s AI – if that’s the route they go down – and developing bots on the back of their systems. But the larger plan for Facebook may be simply to get businesses more reliant on their platforms.
In an interview with Wired back in October, David Marcus said that advertising was where Messenger would make their money.
"When you're a business that generates most of its revenues from advertising, it's just a better business. eBay takes a cut of every transaction and listing; Alibaba does all that for free, and makes money from advertising. Alibaba is bigger than eBay and Amazon combined, and is growing much faster. We take the same approach. We want the maximum number of transactions on the platform, while enabling the best possible mobile experience for commerce. The margins on payments aren't that high, and we want the broadest reach. Businesses will want to pay to be featured or promoted - which is a bigger opportunity for us."  
One way they could do this is through featured apps – just recently, Business Insider reported that some users were seeing a suggestions bar of brands you may want to chat with in Messenger, alongside People and Groups.  
In that sense, Facebook could offer Google-style display ads to help promote your business to interested users, based on their Facebook actions and interests. This could also apply in relation to each query – if you type into the search bar that you’re looking for a shoe store nearby, Facebook could charge businesses to be the top match for such a query. And interesting to note on this, Google says interest in “near me” search queries has increased 34x since 2011 and nearly doubled in 2015 over 2014, with the vast majority of those queries (80%) coming via mobile. Of course, being able to pay to rank higher in such searches opens the system up to exploitation, as Google’s found previously, but there may be a way to monetize interest and get businesses to pay for more exposure.
But the real winner for Facebook is likely more users on Messenger. If they can continue to boost user numbers for the app, and introduce new functionality to keep them within the app for longer (and away from competitors like Snapchat), Facebook has more opportunity to monetize that audience and become a more valuable platform for brands. As with most Facebook projects these days, monetization if likely not the main aim at this stage, attention is. Once you capture that, monetization options become easier.
As noted, Facebook's expected to announce their bot store this week, so expect a flurry of announcements around F8 in relation to Messenger bots and the future of Messenger commerce. It may still seem like a way off, but the time where your business will be constructed around smart computer systems may not be as far off as you might think.
There might come a day soon where all you need to do is fill and ship orders, while a computer handles the rest.

No comments:

Post a Comment