Today we expect consumer apps and services to be always available, mobile friendly and highly integrated. We demand a social and seamless experience. This often stands in stark contrast to the business environment: Disconnected applications and manual workflows are still all too common. Virtual assistants or chatbots can help to deliver this efficient experience also in the enterprise world.
The interaction with virtual assistants – or chatbots – usually happens in messenger applications such as Skype, Slack and Facebook Messenger. To provide the most value, bots need to be smart and understand natural language. Then they become part of the team. A human can converse with the virtual assistant much like chatting to another human resulting in a more social experience.
The integration with other systems and data sources is even more important. To provide a seamless experience, bots connect to different back-end systems. The challenge is to hide the complexity and overcome technical limitations of legacy systems. In return, bots can provide tremendous value to businesses. Processes will become more efficient. Automation of more and more tasks means cost savings, less human errors and 24/7 availability. By using artificial intelligence, virtual assistants can learn and even provide new insights to further improve processes. If done right, employee satisfaction will go up due to the seamless and efficient experience.
Understanding natural language and supporting multiple interfaces
One of the key features of smart chatbots is the ability to understand requests that are written or spoken by humans in their language. This means you don’t have to learn a new language or tool in order to send requests to the virtual assistant. You can just send an email, SMS or instant message just like you would send to a real person. Or you can use your voice.
A good example is a virtual assistant for IT support. Support and operations teams typically get requests via email, which is then added to a ticket system. Traditionally this work was done manually. The first step towards automation often means that the content of the email is pasted into the description of the ticket. However, detailed categorization, prioritization and assigning the right contact is often still a manual task. Artificial intelligence can help here. With machine learning the bot can automatically classify tickets and detect similar requests. Similar requests with similar outcomes can be automated completely.
Hiding complexity by focusing on what’s important
Many mission-critical back-end systems are quite complex because they usually try to solve many use cases, even edge cases. This often means that the most common tasks are overly complex for the consumer. Bots see this as an opportunity. Instead, they tend to focus on the most common set of use cases that provide the most value to the consumer. By integrating with different systems and data sources smart bots know about the context and can already make certain assumptions, just like humans do. And as said above, we can communicate with them in natural language. This simplifies the interaction and user experience. Instead of presenting a complex UI with many (most of the time irrelevant) options, it’s up to the consumer to ask for and provide the necessary information.
For example, consider a time tracking application that employees are required to use. Unfortunately, the application doesn’t work in all browsers. It also requires the user to enter many parameters such as user id, order, and suborder number. In practice, however, many of the parameters are often the same and the detail is not needed. A bot can improve this process. Employees can now simply say or type a message like “Book 8 hours today on Sprint 1”. The bot learns from previous entries and is smart enough to figure out the additional required parameters. And if not, it will prompt the user with additional questions to learn and become smarter. The interaction takes place in Skype or Slack, the tools that employees already use for team communication, thus avoiding a context switch and overcoming the technical limitations of the time tracker’s UI. By understanding natural language and hiding unnecessary complexity, the process becomes much more efficient.
Would you like to learn more about our bots?
If you are interested in one of our bots or need a partner in building a custom bot, please contact Arvato Systems.