This behaviour is similar across all customer focused markets, such as Retail and Finance, where the loyalty of the customer depends on the quality of interaction with the business. Interaction with the customer (here the civilian) is also crucial in the Public domain, where trust in the authorities is determined by the speed and precision of the provided feedback. Chatbots offer a new automated channel for conversation with the customer. One single channel, for instance Facebook Messenger, supports the interaction with multiple businesses and government agencies.
The complexity of working with chatbots is the ability to understand the conversation with the customer. Every sentence typed in by the customer has a specific meaning, also known as ‘intent’. For instance, there are many ways to say ‘My bicycle is stolen’, ‘My bike is gone’, ‘They nicked me iron horse’, etc.
The promise of chatbots is enormous. Customers and businesses expect intelligent conversations from day one with interactions in different languages, across many channels and the ability to have self learning capabilities. As always, with hyped new initiatives and functionalities we need to tone down the expectations. A maturity model will help managing the expectations and provide a roadmap that will guide towards an expected result.
In this blog series, we will look at the different capabilities that are required for a chatbot, and then plot those onto a maturity model and roadmap.
There are some critical areas that determine the functionality of a chatbot, Interaction, Intelligence and Integration.
is the area where the end-user experiences the chatbot functionality. Contrary to what an end-user would experience in a Web site or an App, there are no outstanding User Experience features in a chatbot. The communication is done via a command line in tools such as Facebook messenger. The user experience in a chatbot is aimed at facilitating a conversation.
deals with all the capabilities where the conversation is supported by means of intelligence. The capability to understand a sentence and provide an answer most likely to align with the intent of the end-user is what would be the most accurate definition of intelligence.
In order to provide an answer; often the content of the answer needs to be enriched with information from a back-end system. When wanting to know the status of that on-line order via a chatbot, the chatbot should be able to connect with a back-end system to be able to fetch information about that particular order.
The maturity levels can be subdivided into two categories, main levels (1.0, 2.0, 3.0) and sublevels (1.0, 1.1, 1.2,..). The main maturity levels are directly aligned with the Intelligence capabilities. The maturity sublevels are determined by the other two capabilities Interaction and Integration. In the upcoming blogs we look at the different areas that help determine the chatbot maturity.
Speaking of chatbot maturity, a platform that has recently made great strides to facilitate increasing chatbot interactions for its users is Facebook Messenger. Chatbots expose their conversation capabilities through different channels, also known as chatbot messaging platforms. Multiple channels support the chatbot capabilities, such as Facebook messenger, Slack, Alexa and WhatsApp. Since the chatbot capabilities already exist for some years now in many tools, what is the special role of Facebook in Chatbots? The big thing about Facebook Messenger is of course the momentum in the market. In April 2016 Facebook announced “Facebook will now allow businesses to deliver automated customer support, e-commerce guidance, content and interactive experiences through chatbots.”. If Facebook’s goal is to help people communicate, Messenger may be the Facebook-iest product it owns. With an easy to use quickstart you’re up-and-running with chatbots in a quick timeframe. The strong thing about the Facebook messenger bot is the capability to extend the conversation. The ease of building a bot with Facebook messenger, and the additional strong capabilities of extending conversation with images, links and call to action buttons, enabled a quick growth in bots over a short timeframe, with around only 11.000 only in the first three months after the launch. The big question is what the maturity of these chatbots are, and how they meet the expectations of the end-users.
In the next blog we will look at the intelligence capabilities of chatbots.
This blog created with support from Phil Wilkins (senior integration specialist Capgemini UK)