Revolutionizing the Way of Selecting Business Schools (Interview)
Updated: Jul 22, 2021
A pioneering approach to graduate and business school selection combines AI and human expertise.
In the past 10 years, Kalin Yanev has been following his passion for data science and psychology to shape a pioneering approach to connecting MBA and Masters programmes and prospective students. With a PhD degree and a professional background in Organizational Psychology, Kalin has headed Advent Group’s team of MBA and Master’s candidate orientation experts for seven years matching thousands of candidates annually to more than 300 leading business schools and universities.
Since 2018, while Kalin was Advent Group’s Student Marketing and Research Director, the company has been shaping the healthy symbiosis of AI and human expertise that revolutionized the way students and schools connect.
Now business and graduate school prospects are equipped with a unique toolkit to make meaningful data-driven school selection decisions.
At the same time, student recruitment teams are enabled to connect to prospective applicants very likely to commit to their school and then thrive in its unique organisational culture.
How do people make choices and does anything change in their approach when it comes to life-changing decisions such as going to business school?
In contrast to the main trend in decision-making theory which has made several scientists Nobel Prize laureates, I strongly believe that people are not doomed to making irrational decisions. In fact, it is widely agreed that human beings crave logic and rationality for their choices. However, the scientific mainstream has been asserting that, typically, people at best lack enough knowledge and information to follow the well-established mathematical normative strategies or simply tend to obey their political or emotionally flawed reasons. I would not agree that such a statement is necessarily true for the personally consequential case of choosing a degree, on the one hand, and considering the present state of technological development, on the other.
While in the youth of Herbert Simon and Daniel Kahneman it was indeed impossible to know and correctly evaluate all possible alternatives constituting a potential choice, (un)fortunately now we have all the technology we might need to achieve a well-informed decision – especially with the highly potent and flourishing new generation of AI being included in the toolbox.
Add to the equation the extremely high importance of choosing the right degree for personal development and you have all the necessary prerequisites for a rational, pragmatic decision – which many MBA candidates are indeed able to make.
An important point is that it could be very time-consuming and sometimes even financially demanding to identify your best business school from the variety and multitude of options in the world. Therefore, according to both decision-making theory and our experience, candidates who usually are trying to make a rational choice are also greatly concerned about how much time the process of choosing would take.
In summary, deciding which business school to apply to could be a purely rational process in which the evaluation of the effort needed to make a decision plays an important role.
Whom should MBA seekers trust, when and why – humans or AI?
The topic of trust in AI is very interesting in the context of the present state of AI technological development and societal endorsement. I believe that we are on the very verge between trusting AI and an avalanche of disbelief. At the moment the majority of the AI proponents have the same narrative: AI is giving opportunities to marginalized people; it is not substituting for humans but just augmenting their work; its mistakes are not much different to or more frequent than human mistakes, etc. Soon the public will become bored and this is just one step away from losing trust in AI. When distrust happens, a very interesting co-evolution will start. In my opinion, the AI industry has had it easy so far – the public has been mainly either scared or amazed. Both these reactions are based on trust in the technology. When there is trust it is very easy to improve your AI prediction of the behavior of the trusting one because both sides are walking towards each other. You might be tempted to use a pejorative simplifying phrase like “self-fulfilling prophecy”.
On the other hand, in the case of distrust, a predator–prey chase will start. In biology they call it the Red Queen's race. Let's imagine that AI predicts which school I will apply to but I do not trust its prediction. So, when seeing the prediction, I would do something that definitely does not align with the prediction – I would certainly choose some other school. AI in turn could easily predict this new opposing behavior. However, by definition, if it shows me the new prediction, or any prediction, I would simply make it incorrect by doing something contradictory. Naturally, at this point the AI prediction would start to be hidden, just guiding some invisible actions (e.g. guiding a consultant). In their turn, the users would start guessing at what is hidden... And so forth. Eventually, by the logic of the bi-logical hypothesis, the prediction would achieve a stable level of accuracy while both the human behavior which is predicted and the AI algorithms would greatly change.
Well is that not exactly the same as human coaches and advisors? No, because we differentiate between them and while we might not trust one expert, this does not mean that we distrust all of them. In the case of AI, we are still in the phase of just two perceived groups of entities: AI vs. non-AI.
Hence, my advice would be: differentiate between AI technologies. There are good ones and bad ones. A good AI is the one that is able to adequately reflect the phenomenon which it is trying to predict. For a changing phenomenon like MBA preferences, the good AI would be the one coevolving with its targeted social construct. It is not an easy task. It was not easy for us either. So, trust the good AI.
What is unique about Advent Group’s latest brainchild, Unimy, and how is Advent’s 15 years of experience matching MBA seekers to the right business schools integrated in this pioneering platform?
Unimy visitors have access to all accredited business schools on a single platform. A unique AI-powered tool predicts whether a candidate would tend to apply to a given school, taking into account more than 400 factors. The Cultural Fit Test, based on a proprietary Cultural Fit Map, is another one-of-a-kind feature. A very important aspect of the Unimy approach is the human communication. Orientation experts complement or supplement the user experience with the tools, and personal conversations with MBA alumni help visitors get connected and experience the MBA community for themselves.
The words “match” and “fit” both suggest “the right options”. What is the difference between the AI-powered MBA matching feature of the platform and its business school culture fit?
The AI matching would answer the question: “Which schools am I likely to apply to?”, or in other words “Which schools would I like, before applying?” The Cultural Fit test, on the other hand, would answer the question “Which schools would I like, after spending considerable time as their student and being able to immerse myself in their culture?” In a series of interviews with MBA alumni, we have found out that a mismatch between your personal beliefs and a school’s culture may result in serious dissatisfaction with the studies and even dropping out.
Business schools rely on rankings to see how they are doing in relation to their peers – but rankings do not directly measure culture.
How does the MBA matching work? Is it based only on the information provided by the MBA seeker or does it take into consideration other factors?
The AI matching works with all the information provided by the user, which is 29 fields translated into more than 400 features. This includes the MBA preferences of the candidate, but it uses them only as possible factors, not ones that necessarily correlate to preferring, or not, a single school. We also take into account the candidate’s other individual features such as work background, education, etc. The models are unique to a school, so for one school a particular feature might be taken into consideration, but for another, it might be considered irrelevant.
As a result, each user gets a unique list of 10 programs to which they are most likely to apply. They are listed in alphabetical order.
What weight should MBA seekers give to the results in Unimy Match and Fit?
Just as people arrive at a final decision through different routes, each of Unimy’s features adds information for the candidate’s rational choice.
Some have a very clear idea about their limitations and criteria. In this case, we would recommend using the Browse option with its rich set of filters. Other people may want to compare themselves with the rest of the candidates by using, so to speak, the collective knowledge. For them, I would recommend AI Matching. For some people the culture within the school such as the appropriate code of behavior might be irrelevant, while for some this could turn out to be an additional important factor. Some candidates may trust a human expert more than automated tools. Some may want to hear first-hand impressions from alumni.
Do any of the Unimy tools help MBA prospects estimate their chances of admission to the best-matching schools?
We give a general filter on the AI tool according to basic eligibility requirements of the schools such as work experience and managerial experience. We intentionally avoided going into more depth with such an evaluation as being admitted to a business school depends on a complex set of factors.
This interview was originally published in the Access MBA, EMBA and Masters Guide.