Facebook Messenger is becoming a force in the enterprise virtual digital assistant (VDA) space as a host platform/channel. However, the much-publicized number of 100,000 chatbots on the platform is misleading. Most of the current chatbots on Messenger do not fall under Tractica’s four key enterprise use cases (customer service & marketing, e-commerce & sales, healthcare, and business applications), but Tractica does believe that, over the next few years, the platform will be a significant channel for most of the enterprise use cases.
Compelling for both enterprises and consumers is reach and engagement. In terms of reach, Facebook reported, as of September 2017, that there were 1.3 billion monthly active users (MAUs) on Facebook Messenger. The company estimates that 65 million businesses are active on Facebook, and that 80% of those businesses use messaging to reach users. So, let us consider the potential of WhatsApp. While today, WhatsApp is not open to chatbots, there is discussion that the new standalone app for business accounts will reconsider. Eventually, over the next few years, Tractica believes that WhatsApp will host chatbots, adding another 1.3 billion MAUs to Facebook’s reach, with relatively little overlap with Facebook Messenger.
Reducing Friction in Customer Experience
In April 2017, Facebook launched a major update to Messenger, dubbed Messenger 2.0, which contained multiple initiatives to improve chatbot performance and engagement:
- Natural language processing (NLP) expanded to 10 new languages, bringing the total of languages to 21.
- The launch of Discover, which according to the Facebook developer blog, is “a new surface to help people intuitively browse and find the best bots, places and businesses in Messenger. Discover showcases popular bots in our ecosystem, nearby places that people can message, and businesses that offer support or answer questions.”
- Chat extension will allow bots to be introduced contextually to a conversation. This is a key development, as the Messenger user does not have to start their business conversation with the bot, just the business itself.
- Hand-over protocol will allow multiple vendors to create experiences within a single bot. For example, this will allow companies to integrate expertise from a customer service chatbot specialist and a separate e-commerce transaction specialist. Facebook notes in the blog post that this functionality is in limited release.
Another compelling driver for Facebook chatbots is the broad concept of frictionless commerce. In the case of e-commerce & sales VDAs, consumers conceivably would not need to leave Messenger to complete a transaction.
The Converse.AI Case for Improving Work
Facebook will aggressively pursue the business application VDA use case (productivity, collaboration, and workflow/project management) through Workplace by Facebook. Launched in October 2016, Workplace by Facebook is used by more than 14,000 companies (as of September 2017), including Walmart. Chatbots are welcome on the platform. Converse.AI (which recently announced it is being acquired by Smartsheet) has focused a significant amount of its energy on workflow management chatbots with Workplace by Facebook.
When it comes to enterprise chatbots, two tech giants, Google and Facebook, are leaning on small startup partner, Converse.AI. In May at Google.io, Converse.AI launched its partnership with Google to power chatbots on Google Assistant and announced it was a launch partner for Workplace by Facebook. Converse.AI chatbots are designed from the outset to integrate with the existing tools, services, and business workflow of the companies using them, without the need for additional code or software.
In 2015, Chief Executive Officer (CEO) Tony Lucas observed that messaging was getting bigger and bigger, further explaining “… but enterprise use of messaging was incredibly nascent. Humans want to interact with companies via messaging, and automation, in my view, is the only way to make that work.” Thus Converse.AI was launched in 2016 with the goal of focusing on enterprise chatbots.
Lucas built several bots from scratch in 2015. “I started thinking about the use cases, and the challenges, then started to think about the architecture,” said Lucas. “I wanted to take the code out so that HR people or some other non-tech could build a bot. Which brought us to a graphical platform to build these workflows but with a conversation interface. Every time we saw a trend in that, we built a module that could address the issue. We now have 300 modules. For enterprise, 2017 is year of less hype for chatbots and more about, ‘oh yeah maybe this is useful’.”
Addressing the Discoverability Factor
In terms of challenges, the primary issue for Facebook and other messaging platforms as channels for enterprise VDAs is discoverability. For most enterprise VDA use cases, messaging platforms are not yet the most logical place for consumers to look to start an interaction with a company; websites are. Despite Facebook’s new upgraded capability for chat extension, the challenge is a broad one, and it will take several years for consumers to shift that thinking in most markets. Facebook understands this challenge and has doubled down, taking a significant step with the launch of a private beta announced in early November 2017 of its Customer Chat Plugin. The plugin allows developers to integrate Facebook Messenger bots directly into a company website. Additionally, the plugin loads recent chat history for that individual and enterprise, and the history will follow across channels back to the Messenger app. According to the company blog, there are limitations currently for the plugin, as it will not work with Internet Explorer, nor with “in-app browsers on mobile,” which essentially means it will not work within dedicated mobile apps.
Promising Developments for Messenger Platforms
Given this focus on chatbot improvement, coupled with unprecedented reach, Tractica believes that Facebook’s Messenger platforms will become the key enterprise VDA platform over the next few years for customer service & marketing VDAs, e-commerce & sales VDAs, business application VDAs, and, eventually, healthcare VDAs.