In which language is IBM Watson programmed

Artificial intelligence: IBM Watson in a project at Capgemini

As an expert in artificial intelligence (AI), Alexander Hesseler uses, among other things, IBM Watson in his projects in the field of digital customer experience. In an interview, he reveals more about so-called "cognitive services", how an AI project at Capgemini works, and how he introduces students to the topic.

Alexander, how did you come to Capgemini and what is your area of ​​activity?

I am a Principal Consultant in the Digital Customer Experience department at Capgemini in Munich. I started seven years ago to develop the “connected vehicle” topic at Capgemini. Since February 2017 I have been leading a team that uses "Cognitive Services" as part of a partnership with IBM to make our customers' processes more efficient or to develop new business models.

Cognitive services may not be familiar to everyone. What can we imagine by that?

Cognitive services are based on the technology of cognitive computing, or in other words: on artificial intelligence. In our case we are working with IBM Watson. But artificial intelligence is a very broad term. We are primarily concerned with the analysis of so-called unstructured data. This is data, the content of which is not already neatly available, but must first be extracted - for example from images or sounds. If this succeeds, the system can process this data further. An example: We can train the AI ​​to recognize cars on images, e.g. for police searches. The recognition and processing of human language by machines is also made possible in this way: Spoken language is first converted into a text, which the system then analyzes. This is the basis for programming chatbots, for example.

What does working with AI look like in your job? How does a typical project work?

Most of the time, customers come to us with a problem and want to know whether there is a solution with the help of artificial intelligence. An example: A customer from the automotive industry wants to automate the quality check at the end of the painting process, i.e. to detect scratches and paint defects. For a human, this is very tiring work that a computer could do better. In the first step, we therefore test how a possible image recording can take place on site, i.e. which cameras and image angles we best use so that we receive recordings that can be fed to the AI ​​system. Then we train the system to independently identify paint defects and check how high the hit rate is or whether this is sufficient for productive use.

W.ow is the team structured that works on such a project? A lot of competencies have to be covered here.

We use different experts for every step. For the first step, the recording of data e.g. B. about sensors, we need technical experts who are familiar with sensors, circuits and amplifiers. The next step, data analysis, is more about the IT aspect; This is where colleagues from our Digital Customer Experience and Insights & Data department join us. But colleagues from the consulting area also help by designing and calculating a business case from the customer request. It really is an interdisciplinary team effort.

W.Which applicants are you looking for in your area of ​​digital customer experience? What should the perfect candidate bring with them?

We are looking for young colleagues who are passionate about digital innovation; Digital natives who want truly groundbreaking IT services. Ideally, these are people with programming skills - object orientation is important, especially Java and Python are popular languages. It is also a must to be technically up-to-date. It is also very important to be able to work and program in a team: AI is not a one-man show. And if someone can communicate with software developers on the technical level as well as with customers on the business level, then we are very interested.

You have already conducted workshops on artificial intelligence and chatbots for students. What did the participants learn from it?

In workshops like this, we first look at how a chatbot works and how to get it to understand what a user's intention or question is. To do this, we have to teach him to differentiate between different intentions of the user - e.g. a greeting from an intention to order or a change of contact data - and then to help the user accordingly.

In the practical part of our workshops, we then develop chatbots ourselves. The workshop groups are then usually free to come up with a topic. For example, last year a group came up with an idea for a chatbot that could find the perfect Christmas present. Another should analyze disease symptoms. And that the teams then present and pitch their results afterwards is also part of it and is often very entertaining.

If you also want to get a taste of the latest trends in a workshop, you will find all current dates on our event pageYou want to find out more about job opportunities in the digital customer experienceExperienced? Take a look at our current job advertisements. Feel free to contact us if you have any questions.