David Yang: “We need to learn a new open-mindedness”
- David, the business world knows a wide range of your entrepreneurial interests. However, your most striking projects are still connected with the IT industry. In this regard, there are two questions: why did IT become your priority, and what prompted you to create the YVA platform?
In regards to IT, I think, I can say it was a kind of accident, at that moment. I've wanted to be a physicist all my life, and the first time we did LINGVO, I wasn't going to give up science. I even hid from my parents for a long time that I had left physics. So I started doing IT almost by accident, but not by chance I continued, because technology is an area that can change the world across borders, across huge distances. Back then, before the Internet, it wasn't so obvious. Our programs were distributed on five-inch floppy disks. Now, of course, it has become clear that in general, any business should be technological in principle: industrial production, agriculture, and the service sector — all these businesses should not just become technological; they should become businesses built using the most modern achievements of artificial intelligence.
As for YVA, this is an attempt to apply ABBYY's many years of experience in the field of artificial intelligence to corporate relations, to DIGITAL RELATIONSHIP, that is, to digital communications in organizations. We believe that the knowledge and wisdom of the organization lies somewhere in a very complex segment, like digital communications, and there is reason to believe that we were not mistaken. As a result of communication with clients, we finally realized what value we have created. We help organizations achieve greater efficiency by applying best practices. YVA is an artificial intelligence technology that accumulates best practices and helps employees to apply them; and in doing so, they observe and identify customer problems in the very beginning and prevent them from becoming large and difficult to solve.
- What is the mechanism of its work? For example, how will it bring efficiency to business in personnel policy?
The YVA platform is a technology called predictive behavioral Analytics, and we have built several solutions on this platform. I will briefly describe three of them.
The first solution allows you to identify burnout employees and understand its causes. Burnout is a very painful phenomenon. Occupational burnout has been listed by the world health organization as an official disease. According to various data, about 40% of all employees are in some degree of burnout, and this, of course, significantly reduces the effectiveness of the organization, its ability to develop and achieve new goals. Identifying the causes of burnout is a strong reason to improve the situation in the company, to create an environment favorable for professional growth of employees, motivating them to be creative. This solution is called "Burnout detector".
- By what criteria is burnout detected?
As a source signal, the method "Employee communication data" is used, that is, who, when, and to whom writes a message. Without opening the letter envelope. To detect burnout, the system does not need to look at the contents of the correspondence. There is an analysis of tens of thousands of data indicators for each employee. The neural network observes a very complex signal and changes in this signal and thus learns to detect the phenomenon of burnout and ultimately predict dismissal. Simply put, the system provided information on all employees who left and gave the opportunity to analyze all the data for a year or two before that, and the system using unsupervised learning methods, found the changes preceding the fact of dismissal. Starting to analyze it, we saw that sometimes during burnout the person is a little reduced length of working time, that is, the first letter he sends to 15-16 minutes later than usual, sometimes it's a bit slower responding to letters, different builds communication with management, with their subordinates, external contacts, etc. A complex phenomenon is almost impossible to see, but the neural network is trained to see the changes that signal the burnout and subsequently lead to the dismissal of the employee.
- And what else?
The second solution is Leadership Analytics. This technology is about leadership styles. Applying it, you can, in particular, determine which leadership style gravitate current leaders. The system identifies 6 types of leaders. One of the most problematic types is the detouched leader. Accordingly, this is a person who is not focused on people or on the result. There are also types of democratic and authoritarian. Leadership Analytics helps organizations understand which leadership style is best applied to them.
This is the first part of the solution. The next step is to identify Natural Leaders or informal leaders of the company who do not have the status of a leader, but who manifest themselves as leaders. This is very important for managing changes in the organization, for managing the transformation of the organization, for the formation of the so-called personnel reserve or High Potentials. These are the people you need to rely on to make something new. For example, we found that groups run by detouched leaders had worse performance than others. Burnout is higher, turnover is higher, happiness - lower, and so on. We realized that the problem is in the rejection of a conniving leader, it is clear that it should be replaced. But who? Where do I get a replacement? And here the system immediately offers a number of options from among the most natural leaders. Those to whom, in fact, people are drawn, to whom they go for professional or worldly advice.
Or another example: the company aims to create an innovation track, a division that deals with innovative technologies, which will also have its own services in different regions. And it immediately raises the question of which of the 20,000 employees listed in the company's staff, to appoint the head of this division. Using the capabilities of YVA and its Leadership Analytics solutions, you can access the list of natural leaders of the company that are more open to innovation at the touch of a button.
The third solution is Corporate Agility Analytics. This is a very important point in mergers, acquisitions, changes in the organization, when it is necessary to move away from outdated, inefficient approaches to making management decisions that require each decision to be coordinated with management. This approach slows down business very much. Modern business must rely on Self-Empowered employees, that is, on the one hand, initiative, and on the other hand, equipped with the freedom to set their own goals. Imagine a product-manager, developer, testers, marketing, technical writers self-organized into a working group, created a given product, got a happy customer audience (customer success) and earned additional revenue. Just imagine: themselves, without millions of approvals from management or detailed instructions. This is the dream of modern companies. But how do you find such people in your organization? How can best practices be extended to lagging groups? YVA allows you to "see" all such groups in real time, to understand how flexible they are, how capable of self-organization, how quickly they make decisions without top management. YVA automatically detects such dynamic cross-functional groups and eventually produces a figure that we call the YVA Agility Index. You can monitor the change in this index. If management has made a decision to become a dynamic company in two years, it will see how this level, the index of dynamism, increases monthly.
- Are there already examples of practical application of YVA and its individual solutions?
Yes, there is an Employee Retention Solution that was presented in the beginning of this year. And we already have more than 40 clients, with a staff of several employees to two hundred thousand or more people. So YVA is now analyzing billions of accurate data. At the same time, we do not touch the data ourselves, except for those situations when our design clients send us anonymized data for the purpose of training neural networks. In other cases, the data remains in the perimeter of organizations due to security and confidentiality requirements.
- Nowadays, many machine-building enterprises that implement digital transformation are faced with the problem of too narrow horizons of engineering and management personnel. After all, now technical specialists, IT engineers, programmers already need a certain level of cross-knowledge in many areas, for example, in biology, chemistry, etc. Have you ever thought about creating a neural network that can train specialists in some related disciplines that are not directly related to their profile education?
I find difficulty in giving an example when AI technologies are directly engaged in training, although there are certainly such solutions. In the case of YVA, we identify growth points, what can be improved, recommend what should be learned. If you are an employee of a company that uses our program, you can log in to your personal account and see your "spider diagram". You can find out that you have certain leadership qualities and get advice on how you can reveal them. For example, the manager may see that he praises and encourages his subordinates less than is usually accepted. Or the system will draw his attention to the fact that he has weak presentation skills, and tell you how to tighten them. In general, the system monitors a large number of skills. We will have a Market Place at some point, and everyone in the personal account will not only be able to identify the skill that is missing, which needs to be improved, but also use the option "get an online course". And there are a huge number of organizations that are engaged in corporate training, create special training videos, and so on. That is, the organization can train people to go further in the direction in which it considers necessary. We just tell you what to learn.
- You mentioned that YVA does not directly analyze the content of emails, however, it processes personal data, which is its "breeding ground". In this regard, how to avoid difficulties with compliance with the law and with the ethical side of the case?
YVA does not require analysis of the content of correspondence for the purpose of identifying employee burnout. But, for example, to identify leadership styles or Corporate Agility, you need to give it access to "open the envelope" of the letter. What does she do after opening the envelope? Identifies some semantic tags: praise, problem statement, conflict, and so on. She writes some labels on the envelope and closes the envelope back. YVA does not store the contents of the correspondence. However, this is personal data, and every country has legislation related to it. For example, in Europe, an employee needs to get permission for analysis. In some countries, it is sufficient to notify employees of the fact of the analysis for such a purpose, in others-it is necessary to obtain their consent.
So employees are either, at a minimum, notified, or have given explicit consent to this analysis. In return, they get a personal account that allows them to improve their performance and, as a result, more likely to get a promotion. It's very simple. If you want to advance in the service faster, then you certainly need to know the points of growth. So this is the stage of exchanging values. Now, let`s talk about ethical aspects. The system does not analyze personal data sources, it does not analyze personal messengers, social networks — all this is absolutely forbidden. It analyzes only corporate data sources that belong to the organization. Again, the person consents to the analysis in order to improve efficiency and to make their organization a better place to work. In other words, our ultimate goal is to make each employee more efficient.
- So there is no suspicion of the invisible presence of "big brother" raises, isn`t there?
I`m afraid, you are wrong. Any Analytics is to some extent an invasion of human space. A person by nature always wants to protect their privacy. But the question is, for what purpose? In the case of our system, employees know that YVA is their assistant, assistant, and coach. And the results of the analysis, depending on the same standards of the organization, are usually completely anonymized and provided for an average subgroup of at least 10 people. And this preserves complete privacy for employees. Or this organization offers its employees some other level of privacy.
- Does YVA have any analogs, competing solutions from other developers?
There are several solutions that deal with predictive behavioral analytics. Among them, for example, a system called Humanize, and Microsoft is preparing a product called Workplace Analytics, which analyzes corporate communications of employees in the mail. But we haven't seen a single solution that could demonstrate employee burnout. We have not seen solutions that show us leadership styles, reveal hidden opinion leaders. Our difference is that we combine so-called "passive" and "active" Analytics in a unique way. Passive Analytics is the Analytics of objective data and digital communications. Active Analytics is a quiz. YVA has built-in special survey systems that are controlled by AI. Each employee receives a sixty-second small questionnaire once a week. And the system asks each person its own questions. She distinguishes employees, knows the characteristics of their characters and gives each such question that is right now for him. For example, " Peter, you have three extra tickets to the cinema, which of your colleagues would you give them?" This combination of active and passive Analytics gives a completely unique result.
- Is YVA a solution for office employees? And can it be used, for example, in production in engineering companies?
Line employees in production today leave almost no digital footprint that YVA could use. But next year we will present technologies that allow us to involve line staff in production, transport, and retail. This is a big interesting topic.
- A number of universities are now forming AI departments. How much, in your opinion, will this actually help to strengthen the personnel base in your segment?
I think it will help a lot. A few years ago, we opened two departments at the faculty of computer science and high technology: the Department of Image Recognition and Text Processing and the Department of Computer Linguistics. In addition, we recently opened the ABBYY Labs laboratory for computer vision and natural language analysis at MIPT. This is very important, it is simply impossible without it. The market needs experts in the field of Data Science & AI. They are now so popular that we and many other organizations are preparing them from the student's bench.
- That is, universities will not pull such an education on their own, will they? Is it necessary to attract private resources?
It is necessary to attract expertise and real production experience! This has always been the case at MIPT. The Institute was created by Nobel laureates Lev Landau, and Sergey Kapitsa, by analogy with the world's leading universities. It had the original link to the final customer. Until recently, these customers were scientific institutions. At the Institute, the first and second year classes began on the so-called base, and students became physicists with real laboratories, with a real future and real experiments. This was a fundamental difference, and now this tradition continues; ABBYY, Yandex, Sberbank and others have now been added to the basic organizations.
After all, the student cannot be taught something in the abstract, it is meaningless. The modern world is changing so quickly and dramatically that what you are learning now will become useless by the end of your studies. Now you need to learn exactly the neural networks, on the data, on the user cases, on which you will continue to work.
- Russia has created an infrastructure to support startups: there are Skolkovo, IIDF and other institutions. However, we do not see startups playing any significant role in the development of high-tech segments of the domestic economy. There are, of course, breakthrough projects among them, but these are rather isolated cases. What do you think is the reason for this?
I would like to mention two aspects that could significantly strengthen the innovative atmosphere in Russia.
First, it is the image of business as a whole. Instead of bringing criminal cases in the course of economic disputes against innovative entrepreneurs and investors, it is necessary, on the contrary, to promote technological entrepreneurship and bring to our compatriots the idea that the future of our country (and indeed any country in the era of the fourth industrial revolution) in high technology, and therefore in scientists, engineers and businessmen.
We need our Elon Musk, Bill Gates, Jeff Bezos, and other leaders who prove that we can change this world.
Second, it is openness to the world, not isolation.
I live half in Silicon Valley, half in Moscow, and I see the situation from the inside and there.
Silicon Valley is a place where about 80 % of the residents are foreigners and their descendants in the first and second generations. This is about two million people from India, almost the same number - from Asia, a huge number of people from the Middle East, from Europe, several hundred thousand from Russia, Ukraine and the post-Soviet space. Professional and initiative people from different countries come here. They create the same innovations, they exchange cultures, ideas.
This "melting pot", which does not exist in Russia anymore. Something similar was in the USSR, where 15 different cultures were collected. They were united in scientific centers-in Zelenograd, Chernogolovka, Pushchino and Novosibirsk. There were a lot of different people. The same MIPT has always been a very international place. And now how many startups can you name from Skolkovo that would be created, for example, by Chinese groups or international groups from China, Russia and France? I don't know any of them. That's what we need to do.
Skolkovo or Innopolis are the most beautiful projects. I have always been a great supporter and supported Skolkovo, Innopolis and other innovation centers. But I've always said that's not enough.
We need to create an open atmosphere where talented people, engineers, scientists, entrepreneurs and investors from China, India, Pakistan, France and Germany will dream of getting to Skolkovo instead of Silicon Valley and feel safe here. That's when everything will change, that's when it will really be very cool. We need to learn a new openness.