top of page
  • Writer's pictureSlashData Team

Beyond the ‘chatbot’ – The messaging quadrant

These days everyone and their uncle is talking about chatbots as the next thing after apps. But ‘Chatbot’ is a rather poor name for explaining the changes taking hold in mobile. ‘Chatbots’ are a mental model suited to developers. The term means very little to users. Besides, ‘chatbots’ represent only a small part of what is happening in the messaging space.

The messaging bot quadrant

The tech media and blogosphere usually focus on one of three things when talking about ‘chatbots’:

  1. the merits of the conversational UI

  2. the promise and reality of AI-based natural language processing, and

  3. messaging platforms as an alternative to app store distribution.

In reality these three things are quite independent and sometimes the discussion in the media reminds me of the famous Indian proverb about an elephant and the blind men.

The elephant is really about billions of people relying on messaging as their main person-to-person communication tool and in the process getting used to what Dan Grover, WeChat Product Manager, calls the thread-centric UI.

It’s no stretch to see WeChat and its ilk not as SMS replacements but as nascent visions of a mobile OS whose UI paradigm is, rather than rigidly app-centric, thread-centric (and not, strictly speaking, conversation-centric).

The next natural step is to apply the same thread-centric paradigm to communicating with businesses on mobile. It works brilliantly. It works in China, where WeChat/Weixin is on the way to becoming an alternative to the mobile Internet itself. It works in Brazil, where WhatsApp has become a one-stop solution for everyone, from small businesses to government agencies, to manage everything, from transactions to recipes and relationships.

The messaging quadrant

There are thousands of business to consumer chat services already, from early-stage startups like ChatandBook and AndChill, to large corporations like KLM, Disney and NBCUnversal. To compare and contrast all the different approaches we consider two independent axes:

  1. Apps vs. no apps, and

  2. AI vs no AI.

The “apps vs no apps” axis is about whether the service is delivered as a standalone iOS or Android app, or whether the service ‘lives” inside horizontal messaging platforms (Messenger, Slack, Skype, Telegram, Kik, as well as SMS). The “AI vs no AI” axis is about whether the chat service is powered entirely by human operators or uses AI-based natural language processing to automate all or parts of conversations. These 2 axes give us what we call “the messaging quadrant”.

Messaging Quadrant

Apps without AI: There are iOS or Android apps that connect users with human operators through a chat-based interface. For example, HotelTonight offers a human concierge service called Aces inside its mobile app. Snapdeal, Indian e-commerce service, provides chat-based support to its users by connecting them with human operators, again within the Snapdeal app. Yup (formerly MathCrunch) connects students with tutors over a chat-based app. Wellness specialist Vida connects users of its app with human coaches. Pana connects you with a real travel agent, and on and on.

Apps with AI: In this quadrant we find standalone chat apps that use AI to automate all or parts of the interaction with users. For example, GoButler uses AI to help you with travel planning. Ozlo helps you to pick a restaurant or a coffee shop, Lola Travel reinvents personal travel by combining of AI with human travel agents, and so on.

The mobile app approach makes most sense for companies that already have a significant user base (like Snapdeal or HotelTonight). They don’t have to rely on the broken app store distribution to attract users.

For everyone else, I believe a native iOS or Android app is just a stop gap. The high cost of making, distributing and supporting native apps combined with the distribution power of messaging platforms will push the majority of companies to skip mobile apps completely and deliver their services directly inside messaging platforms. Let’s look at some examples. Operator, a US-based shopping assistant that connects users with shopping experts initially started as an iOS mobile app, but recently started offering its services over Facebook Messenger.

No apps, no AI: In this quadrant we find services connecting users with human operators inside popular messaging apps or SMS. For example, Rogers, a Canadian telecom operator offer access to its support operators through Facebook Messenger. Examples are numerous. If this sounds boring, consider the business potential of replacing millions of 1-800 customer support numbers with 21st century technology and user experience. Just replacing clunky IVR with simple messaging plus custom buttons and asynchronous communication with the agents are huge steps forward. (I hope you’re not listening to some boring hold music over the telephone while reading this post.)

There is also a growing number of services connecting users with human operators and providers over messaging platforms. For example, HealthTap makes its network of U.S. physicians available to Messenger users worldwide.

AI without apps: This area generates most excitement with entrepreneurs. It allows access to users without the costs and headaches of native mobile apps, while promising to be much more scalable than services dependent on human operators. AI can have enormous impact without reaching full automation. Consider that in customer support recognising and automatically responding to the top 10 simple queries, while routing everything else to a human operator might allow a big corporation to cut lots of call centre jobs.

Facebook itself is experimenting with M on Messenger, where it uses humans to train AI and create a fully automated personal assistant. Assist uses natural language processing to connect users with a host of on-demand services over SMS, Messenger, Slack, Telegram and Kik. Meekan automates meeting scheduling for Slack teams and on the way helps you to find the flight. Databot allows you to “speak with your database” simply in a Slack chat. The list of automated chat services grows every hour.

One critically important question remains unanswered: How users will discover thousands and potentially millions of 3rd party services available inside messaging platforms? Slack Bot Directory, for example, adopts an approach similar to mobile app stores. Users can browse selection of Slack bots organized by categories or popularity. Facebook and Microsoft seem ready to take a different approach: Facebook M virtual assistant will be able to recommend chat services based on the current conversation and requests. Microsoft demonstrated how the company’s Cortana virtual assistant automatically summons and dismisses 3rd party chat services in a conversation.

A post-app ecosystem in the making

Stan Chudnovsky, Head of Product, Facebook Messenger said at TechCrunch Disrupt that “tens of thousands” of developers are making messaging bots for the Facebook Messenger platform. Already tens of companies are building tools for bot developers – From giants like Facebook and Microsoft, to startups like Smooch, Kasisto, Reply, Meya, and Init.

It’s clearly the beginning of a post-app wave and the hottest opportunity for developers disillusioned by mobile apps. We cannot wait to share the results of our 12th Developer Economics survey, which includes questions on bot development.


bottom of page