What does Ingo have to do with Conversational AI?

Ingo Steinkellner is Chief Technology Officer (CTO) at the Zurich-based software developer aiaibot. He is therefore responsible for everything technical there. He takes care of development on the one hand and the interface to the business on the other, i.e. everything technical for customer projects.

What is AI?

Artificial intelligence (AI) is the attempt to imitate human behavior. One example of this is self-driving or semi-autonomous cars. In this case, AI uses sensors and software to help people imitate how they drive a car. In this process, various factors are processed in order to make decisions in real time. This technology also has areas of application in connection with conversational AI and chatbots. In this context, AI recognizes what people want from it and how it can process and respond to these requests.

What is Conversational AI?

Conversational AI is the application of AI to conversations. This can take various forms. Prominent examples of this are systems such as Google Assistant or Siri. Here, people talk to a machine, which tries to recognize and process what is said and formulate a spoken or written response. aiaibot, on the other hand, works textually with chatbots. People enter their concerns on their computer or mobile device, the machine tries to categorize them and generate an answer or solution.

What are the benefits of Conversational AI?

Many companies have typical concerns and frequently asked questions from customers or employees. This can be automated relatively easily with the technology available today. On the one hand in dialog, on the other in case processing, i.e. answers and solutions that are generated and delivered directly. This application generates great added value for customers using a relatively inexpensive technology. In addition, Conversational AI provides an intelligent assistant and employee around the clock. In addition, there are many digital touchpoints that can be used with conversational AI, for example homepages or additional communication channels such as Slack or Microsoft Teams. Such channels can then be used to automate responses to customer or employee concerns.

Rule-based chatbots vs. conversational AI

At this point, the question arises, what is the difference between rule-based chatbots and conversational AI?

Rule-based chatbots directly suggest possible answers or questions to users. Therefore, rule-based chatbots are not the same as conversational AI, which is as open as possible and allows the user to express their concerns. However, a combination of conversational AI and rule-based is often recommended for chatbots, whereby the situation and the type of dialog should always be taken into account. That’s because there are places in a dialog where you should be as open as possible and others where you need to be a little tighter based on the context.

aiaibot started with rule-based chatbots. However, new modules are to be released from September 1: the AI module and Robot. The AI module aims to connect the rule-based bot with AI at appropriate points in the dialog. Robot works via automation and integration. This module allows you to freely automate processes, design the process flow and connect AI with chabots. This turns Robot into an intelligent assistant and data aggregator that can also handle processes and interact with communication channels.

Another aspect of the new modules is that of transparent AI. Many providers who work with AI provide relatively little data on what makes this AI measurable. The goal for the aiaibot module is to be transparent, i.e. to show important key figures in the context of Natural Language Processing, where you can train classifiers with data. The aim is to record transparent key figures for the individual categories that they train. In addition, a Confusion Matrix should be able to generate the optimization potential that exists in a classifier. This testing module can be used to see how the classifiers can be further optimized to improve the customer experience.

Examples of use cases

• Initial question for the chatbot

If the user comes to a homepage and is overwhelmed by all the menu items, then this is the ideal situation to need a bot that works with conversational AI. At this point, he can openly ask the user what his concern is. They can then type it in completely freely. The bot then recognizes this and can send the user on their way, for example to complete their desired product order.

• Training-based bots

If you try to place learning content somewhere, you can ask the user something they have learned. They can type it in freely and conversational AI can be used to analyze afterwards whether their answer is going in the right direction.

• Credit card emergencies

If you have a problem with your credit card, you can solve it directly via the bot or upload invoices if you have problems with billing. Such problems are handled completely automatically by the bot and can be dealt with at any time, even if no customer service is available.

In the context of AI and conversations, the use cases are relatively unlimited. There are now a few that can be used successfully, but new ones are constantly being discovered. This is a hype that is just emerging and has great potential for discovery. That’s why Ingo also spends a lot of time looking at competitor products and takes inspiration and orientation from how others have done things well.

Learnings

1) Artificial intelligence is the attempt to imitate human behavior. In the context of conversational AI and chatbots, AI can recognize what people want from it and how it can process and respond to these requests.

2) Conversational AI is the application of AI to conversations. People enter their concerns on their computer or mobile device and the machine tries to categorize them and generate an answer.

3) A combination of conversational AI and rule-based is often recommended for chatbots, whereby the situation and the type of dialog should always be taken into account.

4) No matter how much AI is used in the chatbot, it is important to think about where you can create added value. There really must be a use case.

It’s best to listen to the podcast episode with Sophie Hundertmark and Ingo Steinkellner for yourself. Have fun![vc_empty_space height=”40px”]

And if you don’t want to miss any more podcast episodes, then sign up for the mailing list at the bottom of this page.