How Facebook's Messenger Bots Work (and Why it Matters to You)
At their recent F8 conference, Facebook introduced their Bots for Messenger platform, a new way for businesses to build their own, interactive, automated response tools for Facebook Messenger. The idea behind bots is relatively simple - you set up a bot service to respond to users on your behalf. So, for example, if you run a restaurant and you know the most commonly asked question is 'what time are you open today?', you can create a bot that'll provide that info via message, quickly and without you having to interact. Such tools can enable significant time savings for business owners, and with almost a billion people now using Messenger every month, it also works in line with the way people are looking to communicate, helping businesses move in trend with user behavior patterns.
Any way you look at it, Bots for Messenger makes sense - that is, of course so long as people actually want to interact with bots. Are users looking to use Messenger to get responses from brands in this way? And if so, how?
To answer this, Facebook recently gave an overview of how Bots for Messenger works, providing in-depth insight into how the process and bot functions are applied.
Here are the key details.
Growth of Messenger
The first point of emphasis in the introduction of Bots for Messenger is the fact that messaging is the number one way people are looking to connect today. In his opening address at F8, Facebook Chief Executive Mark Zuckerberg noted that between Messenger and WhatsApp, Facebook's now processing around 60 billion messages per day - by comparison, at its peak, global SMS volume was around 20 billion messages every 24 hours.
This data is supported by the overall rise in popularity of messaging - a recent study by Pew Research found that Messaging is the dominant communication option among teens, with nothing else even coming close.
But more than that, according to Facebook, people are already communicating with businesses via message more than a billion times per month - a number that's more than doubled over the past year. Such interest further underlines the use case for improved Messenger commerce and connection options and why Facebook is going down this path.
To support this, Facebook built their new Messenger platform to help brands build improved Messenger experiences.
The core of the process operates via Messenger's 'Send and Receive API' which enables businesses to build automated systems, facilitate direct connection with people, or do a combination of both via a range of different options. Using the Send and Receive API, businesses will be able to share a wider range of content via message to help improve communication, including text, images and flexible templates that offer structured text and additional functionality, like 'Call to Action' buttons in stream.
This, according to Facebook, allows a re-imagining of the capacity of bots, as it takes the system beyond a simple text command interface (or a spoken word response, as per Apple's Siri or Microsoft's Cortana).
For example, under the traditional text interaction system, a customer might get a message like this:
By adding the ability to insert Call to Action buttons, the interaction is immediately enhanced.
And by adding an image, it even further improves the process.
Facebook says this process is much better than a similar e-mail notification - if you were to receive an e-mail, you'd have to click on the e-mail, go to the mobile website, log-in, click around, etc. Via message, the interaction is immediate and can happen all within the context of the Messenger thread.
In addition, the process can be even further enhanced by the addition of more product images, accessible via horizontal scroll (the running joke of the presentation was that this was Zuckerberg's regular t-shirt order, hence the lack of variety).
Immediately, from this process, you can see how the Messenger eCommerce experience relates to the new way in which users are looking to connect - it's simple, it's personalized, it's immediate and it's mobile, four key elements required to appeal to the modern consumer. And then, on top of this, there are bots, which take the process to the next level of functionality.
To demonstrate the utility of Messenger bots, Facebook provided three examples of existing bot services and how they operate. The first is fashion retailer Spring.
As you can see in the above image sequence, the user's able to choose which category they want to shop ('shoes'), specify the type of product ('sneakers') and then get a range of examples sent through to browse by scrolling sideways. If none of those take the users liking, they can ask for more examples and the Spring app will send them five more images of sneakers to look through, as per the last image in the sequence. This process is all facilitated via bots - there's no human intervention in the process at this point.
Where a human does get involved is in the next stage.
In this sequence, the user has clicked on the 'Ask a Question' prompt, which then generates a Messenger query. The user's then able to specify their question and the customer service team at Spring is prompted to answer, which they do in the final image. This hybrid human/bot approach will be integral to ensuring the new system can be utilized to best effect - while at the same time, with every interaction, each bot is 'learning' about your individual tastes and preferences. Eventually, as you conduct more business with a brand via this process, the recommendations it provides will become more customized and better matched to what you're most likely to be interested in.
Worth noting, too, in this entire thread, that there's a prominent 'Block' button in the top right to give users control over who can message them and when - this should alleviate some of the accompanying privacy and spam concerns with the new process.
Another bot example Facebook highlights is Poncho, the weather cat, which, as you might expect, provides information about the weather forecast.
The functionality of Poncho is relatively straight-forward, but what's more interesting about the app is how it was built. For example, Poncho's development team used a CMS tool to simulate the end-user experience in order to make it more human-like - so when you send a message to Poncho, there's a delay in response relative to how long that response is, similar to how long it would take a human a moment to write.
Poncho also has an emotional index to score a person's attitude based on the apps understanding of natural language.
In the above example, Poncho's system would decrease this person's attitude score to dynamically adapt its responses and make the interaction more human and personal, based on how they've initially framed their query.
And the last example is CNN's news app, which enables people to stay in touch with the latest news via message.
As per this sequence, via the CNN bot, a user can get a customized feed of the top news stories of the day sent to them via Message. From there, if you've only got a few moments, you can ask for a summary and CNN will give you the lowdown, with a brief description created by CNN's news team.
These three examples highlight some of the possibilities of Messenger bots and the roles they can play in your day-to-day life. And while not everyone will be happy communicating with bots and interacting with systems via Message, it's easy to see how a lot of users will find such options beneficial. And they may also open your mind about the possibilities of the option for further applications, and how they could be used for your own business.
In order to help fuel these new connection processes, Facebook's introducing a range of new ways for brands to highlight their Messenger presence and more easily connect with consumers. First, there'll be an increased focus on the Message button on Facebook Pages, especially via mobile.
There'll also be a Messenger search tool which will make is easy for users to be able to find businesses on the platform.
This could also be a key monetization point for Facebook, with highlighted brands likely to receive more search traffic. There's also improved Messenger buttons which can be inserted onto websites.
And there's the new Messenger codes which can be pasted anywhere for users to scan in and instantly connect to your brand via message.
There's also a new 'Customer Matching' feature for Messenger which enables brands to match customer details they have on file to their respective Messenger user identities to reach them with alerts.
There'll be special policies around the use of this feature to protect users from spam - for example, it'll cost $99 to use customer matching, at least in the initial stages of the roll out.
Overall, you can see how the adoption and use of Facebook Messenger bots is likely to have a significant impact on eCommerce - maybe not straight away, but over time, as more people start to use the options available - and how it's likely to become a more important part of the overall social media marketing puzzle. The tools could also make Facebook a bigger platform for real-time customer service, putting increased pressure on Twitter. Messenger already has a lot more users than Twitter (900m MAU vs 320m) and if users can start to get quick brand responses and interactions via Message, that could reduce the reliance on tweets to generate immediate reaction.
While some have lamented the arrival of Messenger bots, and even questioned whether they'll be used, the overall trend and message adoption rates all point to the fact that more people want to conduct more interactions via message. If brands can deliver immersive, intuitive experiences, like the ones highlighted here, there may come a time where Messenger bots become the way forward for brands.
Follow Andrew Hutchinson on Twitter