Backstage: Peeking into the brains of BSI Studio
Science fiction films like “Metropolis” and “Her” address the relationship between man and machine and fill entire cinemas with it. At BSI, we prefer things to be more tangible: Therefore we treat the issue of machine learning (ML) with the right touch of pragmatism – with a constant focus on the user and customer benefits. Christoph Bräunlich, ML expert at BSI reveals in the following interview what it is that fascinates him about ML and how much intelligence is to be found inside BSI Studio.
What is your favorite film on the topic of man-machine and why?
The various Star Trek films and TV series, especially “The Next Generation" presented an interesting vision of human-computer interaction and artificial intelligence (AI). Many aspects portrayed, such as voice assistants, smartphones and virtual reality, are a reality today.
Artificial intelligence, machine learning, neural networks … Can you shed some light on them?
Real artificial intelligence doesn’t exist yet. AI attempts to simulate “intelligent” behavior with computers. Machine learning refers to a collection of algorithms that are able to make predictions on the basis of experiences in the form of data. Neural networks, in turn, are a group of such algorithms. These take various aspects of biological neural networks as their inspiration.
What opportunities does ML offer companies when it comes to contact with their customers?
With a modern CRM, companies can compile comprehensive data about customers, their communications and behavior. This data is usually so comprehensive that it can no longer be thoroughly analyzed by individuals. ML algorithms can help with the mining of this data in order to reap useful information from it. ML is rather like a data detective, so to speak. For example, it can help marketing specialists and sales reps to identify those customers in danger of jumping – keyword: churn detection. This means increased efficiency and offers unimagined opportunities for shaping customer relationships.
Where do you think the skepticism held by many people about AI comes from?
Many science fiction films and books portray a bleak future: An artificial intelligence attempts to take over the world. In my opinion, the risks are rather with the AI users and developers, because it can be used for dubious and immoral purposes such as monitoring citizens or customers. The ethical questions must also be considered when using ML.
What does ML have to do with the marketing automation platform BSI Studio?
That question is quickly answered: BSI Studio is a digitization platform that combines automation with process optimization. ML takes automation to the next level to support users, whether with a tailor-made customer dia-log or with automated processes. I think this is a must for a modern tool!
Where exactly is ML to be found in BSI Studio, and what is yet to come?
BSI Studio supplies among others different types of neural networks, so-called brains. These are built into (customer) stories. The brains determine, for example, the best time to establish contact based on the habits and preferences of each individual customer or find the right target group for a marketing campaign. Good ideas about brains often also arise among customers. We are only too happy to make use of these.
Does a BSI Studio user have to be able to program, or have ML expertise, in order to profit from the brains?
We have three user groups in focus with BSI Studio:
- Marketing specialists who draft a customer story should be able to choose from a selection of existing brains in the Step Library, and then use them easily.
- A technically sophisticated BSI Studio user can compile Cortexes in a graphical Brain Editor. The editor makes it possible to merge development steps that ordinarily would have to be programmed using drag and drop. An understanding of data types and vectors will be necessary here.
- The data scientist can configure the inner workings of brain blocks, which, depending on the block, requires expertise in neural networks or genetic algorithms.
BSI Studio with ML is also a wonderful tool for a “machine teacher”, because it combines all aspects of the “teaching process” – from compiling data to training and testing, through to publication – in just one interface.
Let’s look into the future: What will be possible?
I would like to see stories that optimize themselves, such as with the goal of boosting sales and thereby profits. This story should be able to make independent decisions in the process that lead to success for as many participants as possible and thereby support the corporate goals. This means no longer working with scores or the like – a brain no longer needs such technical diversions.