ChatBot Basics

More and more customers expect more and more personalized communication via messenger services and can be addressed faster and more specifically in this way. With ChatBots this is possible.

Der Siegeszug der Messenger

Messaging Apps vs. Social in einer Diskussion über Chatbots
Messaging Apps vs. Social

Source: BusinessInsider http://www.businessinsider.de/the-messaging-app-report-2015-11?r=US&IR=T

Messenger services such as WhatsApp and Facebook Messenger have long since overtaken traditional social media platforms in their use. Services and simple questions can also be answered faster and more customer-friendly via Messenger. For a company, however, this is hardly representable in terms of costs and personnel.

The solution is ChatBots, i.e. automated systems that talk to the customer and answer the most important questions in dialog.

What are ChatBots anyway?

A ChatBot is a dialog-based communication interface. Basically, they are nothing more than a full-text search that follows a natural dialogue and is thus considerably simplified for the customer. A simple ChatBot follows a defined dialog scheme and guides the customer to his goal via specific questions and options.

Within a hybrid system ChatBots are used to answer customer queries as far as possible. As soon as they can no longer make any progress themselves, they seamlessly hand over the dialogue to a real employee. He or she will then continue to deal with the request. The customer should not notice at this point that a break has occurred.

The Technology behind the ChatBot

ChatBots are experiencing the big breakthrough because they are increasingly able to integrate technologies such as Natural Language Processing (NLP) & Natural Language Generation (NLG). The customer no longer has to follow a rigid dialog schema, but can formulate his request in natural language, which is then interpreted by ChatBot. The result can also be output in natural language. Such ChatBots become Intelligent Personal Assistants, Siri from Apple or Alexa from Amazon are the best known representatives. NLP is already offered by ChatBot platforms as NLP-as-a-Service and is part of the standard.

More and more often, such core technologies are integrated through communication between ChatBots. A ChatBot takes up the customer’s question and forwards it to a special NLP bot, which interprets it and returns it to ChatBot in a form that it can understand. The latter then transmits the result to the customer. In this way, additional functionalities can be added and search queries and information from other company divisions, remote databases or even external systems can be used.

Software robots, i.e. Robotic Process Automation RPA, mimic human interaction with software interfaces and allow ChatBot to simulate a user making inputs in a system. This enables ChatBot to create and modify records, for example, to take orders or update the customer’s contact information. Some companies are already experimenting with offering their catalogues via ChatBots: the customer is guided through a dialog to the product, ChatBot takes the order and also offers other products that match it.

In the medium term, ChatBots will also have access to more and more elements of artificial intelligence and machine learning. ChatBot learns the preferences and expectations of the customer and can communicate even more naturally and specifically. In addition, the accuracy of search results increases enormously, especially with unstructured data or inaccurate search queries. According to NLP-as-a-Service, cognitive analytics and the warehousing of unstructured data volumes will further advance the topic of ChatBots in the near future. Machine learning mechanisms become more readily available, and this also applies to AI to a limited extent.

ChatBot Applications

ChatBots can be used in various areas: from pure data retrieval and data sharing to marketing and sales, communication and entertainment. All these areas of application make sense both in external communication and within the company itself. A ChatBot-based ordering system can be used, for example, for customer catalogue orders or for employee travel bookings.

Most ChatBots are developed on specialized platforms. Different communication channels such as Facebook Messenger, WhatsApp, Slack etc. often require a specifically adapted ChatBot, since the expectations of the communication behavior are different in each case. The main focus of ChatBot development is no longer on programming skills, but instead on the use case, the planning of dialogs in terms of form, structure and characteristics and the sensible division of data areas into several coherent ChatBots. The underlying data basis and integration into existing communication and CRM systems must be taken into consideration.

How to proceed

Some experts are already predicting the end of the classic website in favor of ChatBots. Even though there are already more than 100.00 ChatBots available for Facebook Messenger today, we think this is a bit exaggerated at the moment. However, we see ChatBots as a very powerful and for the user simple communication medium for various fields of application, which is growing out of its infancy with rapidly increasing speed and is entering our everyday life. Companies and especially the marketing department should now deal with the topic ChatBot and lay the foundations for its use.

Here are a few links to help you get started:

 

  • On BotList you can find and try existing ChatBots
  • ChatFuel and IBM Watson are just two of many different development platforms for ChatBots
  • Interbot provides APIs that allow Bot2Bot communication
  • mycroft.ai is an Open Source Intelligent Personal Assistant, which also works on Raspberry PI and is supposed to build up an open source counterpart to Siri and Alexa.
  • Forrester has created a list of providers for ChatBots in Customer Service, which is summarized in ComputerWoche (German).
  • On Hubspot you will find an excellent overview of the technology and how to develop your first own ChatBot on ChatFuel.

In the context of an ISG ProviderLens I will look at the topic ChatBots in the upcoming Social Business Provider Lens and subject the market and providers to a thorough analysis. Providers are invited to participate in ProviderLens and present their solutions to us.