By: Alain Kaeser, Chief Science Officer and Founder of Yseop
You might be wondering how the idea for Yseop was born. Today, Alain Kaeser, Chief Strategy Officer & inventeur de la technologie of Yseop, shares his personal insight behind his founder’s story of how the company first came to inception.
It all started with a dream…or was it?
Yseop was first born out of a childhood dream of mine which has quickly become a reality over the last quarter of a century. This dream was to create an extraordinary, “magical” toy which did remarkable and unthinkable things, until now.
If this child was born in the 18th or 19th century, perhaps this dream would have been to design a universal clock, or the concept of thermal machine relegating – to the past all that had preceded it!
By the time this dream was embodied, we’re now at the end of the 20th century, whereas the invention of powerful machines often make people dream less; it’s now computers that carry the technological revolution. What once seemed like a “magical” toy will now be a computer. Obviously a computer not quite like others, an extremely intelligent computer envied by all! But what do computers know…then and still today? Meaning, they were originally designed for simple tasks like calculating, sorting, scheduling, and even tasks like drawing a little later.
At first, the success of computers made them indispensable in the economic spheres of everyday life. Since then, computers have taken on a vast set of tasks (formerly only performed by humans), but can ultimately do what they have always done: leveling off the limits of the human mind and amplifying their power. At the dawn of the third millennium, there were still many tasks to which computers could not do: tasks which were not mechanical or sequential but rely on the ability to reason, adapt to changing situations, make decisions, and ultimately communicate. For example, the activities of “white collar workers” are still very manual today. Word processing is great, but it’s still only a slight improvement over using pen and paper. With the disappearance of secretarial work in companies, executives now spend a considerable amount of time writing and generating reports.
It was necessary to take an additional step that researchers partially took in the 1980s by imagining a new generation of computer programs. For instance, decision support systems, expert system programs or artificial intelligence imitating a human’s ability to reason, deduce and advocate. These solutions worked technically, but success was rarely achieved because these tools suffered from two ailments that held back their expansion.
More recently, a new wave of artificial intelligence technology has emerged such as machine learning and deep learning, seeking to achieve the same goals. It is successful and raising a lot of hope, but suffers from the same problem of acceptability and explainability.
First being, low accessibility by users who have trouble accepting a solution which seemed to “reason” for them resulting in difficulty trusting answers that were so mysteriously obtained. Second, these solutions did not fully complete the process and therefore, did not reach sufficient user satisfaction. It’s not enough to simply know what to do – one also needs to understand exactly how to do it.
For example, if I’m a credit analyst to whom my digital decision supports other assistant designates of the approval of my client’s credit application, I still need to be able to justify why the client can be granted this loan. In short, my digital decision needs to be able to effectively communicate. In this situation, the credit support assistant does not perform this task, so the analyst must fend for himself. Yet the analyst doesn’t fully understand the reasoning used to grant his client the loan! Typically, manual tasks require one’s ability to justify and communicate reasoning and intelligent decision-making solutions cannot help with this.
At this point, the child’s dream has become Yseop’s dream: to build an intelligent and interactive solution capable of reproducing and automating a variety of tasks which require communication and thinking only a human can achieve. For this dream to become a reality, we need a solution capable of communicating in natural language with the semantic and linguistic subtlety of a human. Let’s return to the previous example of our credit analyst. Yseop’s solution has the ability to clearly explain the financial analysis and reasoning behind a loan recommendation. Taking this even further, Yseop explains decisions making details in an intelligible manner, by generating automated insights and reports.
Admittedly, the “it was enough to” required almost 25 years of conception, development, industrialization, internationalization, but that is another story…
