Clippy. SmarterChild. Tay AI. If you remember these names, you'll note that chatbots certainly have a checkered past. Despite that, chatbot use cases have a place in the enterprise today.
A chatbot -- or just 'bot' -- is a user interface built on a back-and-forth conversation between a user and an app, likely with the ability to process natural language inputs, and possibly with machine learning or artificial intelligence to help add context to the interaction.
Bots have been the subject of occasional hype over the last few years, but in their core components, they're just like any other type of app or user interface that can bring benefits to business users.
A chatbot application can have varying levels of complexity.
If the user interacts with the bot via voice, then it requires a speech recognition engine. Machine learning has improved the ability of software to translate audio speech into text at a basic level, but a very basic chatbot can even simply use "dumb" voice transcription. That technology transcribes voice into a written script. Text-based bots can skip this step.
The next level is natural language processing. Given all the ways that a user could phrase a single request, natural language processing attempts to deduce the meaning and convert it to an appropriate format to query an application. Finally, more advanced algorithms or machine learning and artificial intelligence can add even more context to a request.
All of these components can come in packages such as libraries that live in client or server applications, or cloud services accessible via APIs.
The back-end integration for a bot looks just like that of any other application. Many of the concepts that organizations put in place for enterprise mobile app development—abstraction layers, mobile backend as a service, and some rapid mobile app development techniques—are very similar for chatbot integration.
On the client side, IT can implement chatbots in mobile apps or on websites, but one of the most defining characteristics of chatbots is that users can interact with them through third-party apps, such as Facebook Messenger, Slack, Twitter or simply mobile SMS.
Build your own bots
In a way, the enterprise has been using chat-like interfaces for decades. Telephone-based interactive voice response systems -- via touch-tone keypads since the 1980s and voice-based interfaces since the 2000s -- are a common example of how an application user interface can provide easy-to-use, back-and-forth conversation.
The big changes of the last few years are simply more sophisticated voice recognition and context, in the forms of natural language processing and artificial intelligence, in addition to new real-time capabilities for the user interface in the form of rapidly-expanding chat and messaging apps.
By keeping these concepts in mind, enterprise IT can keep from getting carried away by any particular wave of chatbot hype. Still, there are plenty of new and interesting chatbot use cases for the enterprise.
In the consumer world, customers can use chatbots to order pizzas, pay bills and talk to customer support. For chatbot use cases in the corporate world, an executive in a board meeting could use an enterprise bot to explore the latest sales numbers just by asking a voice-based digital assistant. An IT administrator could ask a bot in his corporate collaboration platform to check if any servers are experiencing an unusually high resource load. Or frequent business travelers could use bots to classify expenses as they go.
Chatbots are beginning to take advantage of more advanced contextual capabilities, such as the ability to discern the user's mood. For example, a consumer-facing chatbot could offer an unhappy customer a promotional product or discount. This may seem a bit unrelated to employee-facing applications, but undoubtedly, some independent software vendors will look for a way to incorporate such features.
As enterprise devices diversify, so will application user interfaces. The rise of mobile devices and web-based SaaS in the last decade has pushed many IT organizations to get more comfortable offering a variety of ways for users to access applications. Bots are yet one more nuanced option to consider, and appropriate chatbot use cases are now emerging in the enterprise.
- Weighing the Options for an Enterprise Mobility Strategy –SearchSecurity.com