Several recent articles have promoted conversational, text-based UIs as a forthcoming innovation. These UIs promise to provide a natural, comfortable way to access information and services in the enterprise and at home, integrating as necessary with voice interfaces.

Slack, Facebook, and Microsoft have opened up APIs in the hopes that developers will add value to their respective messaging platforms.

Slack API

The Slack API lets you build bot users that “enable teams to conversationally interact with external services or your custom code…” The bot user can

  • be invited to participate in channels,
  • monitor and participate in conversations with other users in channels

Botkit from Howdy.ai is a recommended (open-source) platform for building bot-users. It uses simple keyword pattern-matching to trigger responses.

Facebook Messenger Platform

The Facebook Messenger Platform lets you build bots that interact with users across “all platforms where Messenger exists.” Facebook seems to recognize that Messenger bots will be marketing tools, and specifically mentions support for brand and calls to action.

Furthermore,

The wit.ai Bot Engine effectively turns natural language into structured data as a simple way to manage context and drive conversations based on your business or app’s goals.

As mentioned in a prior post, wit.ai lets a developer classify natural language statements into a set of known intents. For an e-commerce site trying to market using Facebook Messenger Platform, “intents” may be things like

  • Ask if an item is in inventory
  • Inquire about the price of an item
  • Purchase an item listed in inventory
  • Return an item

By allowing numerous natural ways of doing the above, by enabling customization to a particular e-commerce site’s inventory, and by providing a dashboard for continuous improvement of the bot, Facebook hopes Messenger will become a channel for marketing and e-commerce.

Microsoft Skype Bots

The Microsoft Bot Framework lets you “build and connect intelligent bots to interact with your users naturally wherever they are”. Bots are built on top of an open-source Bot Builder SDK.

While Facebook Messenger Platform uses wit.ai to classify natural language statements into known intents, the Microsoft Bot Framework uses the Language Understanding Intelligent Service for the same purpose.

Google

Google has announced an intelligent bot running inside its new messaging app, Allo. However, this seems to be a closed bot, providing information from the knowledge graph Google has already built.

Business Models

The business model considered in the previous post seems to remain valid: Organizations that are already using one of the above messaging platforms might be willing to pay for a bot that provides information and services highly relevant to the enterprise—information and services that would otherwise require costly human mediation or time-consuming search. One example is Wade & Wendy, a conversational recruiting bot. NextIT has been in this space for some time, and butter.ai is providing enterprise file search through a conversational interface. Competitive advantage in such a model will arise from

  • selling capabilities beyond those of the IT departments of organizations,
  • speed and efficiency (cost) of implementation,
  • levels of automation, eliminating human labor,
  • customization to the specific needs of the organization.

This model would also apply to companies that provide a conversational UI to existing complex applications. For example, sudo.ai and Birdly, which provide a conversational UI to Salesforce.com.

Facebook seems to be enabling brand engagement via the Messenger Platform, and the business model here is obviously very marketing oriented. Marketing and IT departments at large companies are likely to project their brands via such a platform, but there will be opportunities for agencies that can build engaging chatbots—much like the opportunities for agencies that can build engaging web sites and mobile applications. Competitive advantage in such a model will arise from

  • selling capabilities beyond those of the IT departments of organizations,
  • quantifiable engagement and sales closure
  • creativity of brand expression

There may be a horizontal technology (for example intent classification from natural language inputs) that becomes competitive to wit.ai and LUIS. However, these are currently free services—their cost is borne by Facebook and Microsoft (respectively) in the interests of promoting adoption of their social media and cloud (respectively) platforms. It will be difficult to monetize a horizontal technology. However, a company that builds one could hope to be acquired by someone whose platform it helps.