MindEscape has moved beyond HR and recruitment into a new frontier: venture capital. The AI-powered serious game developed by SeriousMind is now part of cutting-edge academic research that uses machine learning to predict which startups are most likely to succeed.
The research comes from the doctoral dissertation of Dr. Nikolaos Takas, Senior Partner & CEO of THERMI Venture Capital Management, completed at the Department of Economics of the Aristotle University of Thessaloniki. The dissertation presents a comprehensive AI framework for predicting startup viability—and MindEscape plays a central role in one of its most innovative components.
Perhaps the most original contribution of the dissertation is the use of MindEscape to extract insights about whether founders’ personality traits affect startup viability. MindEscape assesses founders’ personalities through puzzles in a virtual environment, rather than traditional Big Five questionnaires, which are prone to social desirability bias.
Instead of asking founders how they describe themselves, the game records how they actually behave when faced with complex problems under pressure: whether they consistently try different approaches (Openness), whether they follow a structured plan (Conscientiousness), or whether they show signs of stress when facing difficulties (Neuroticism).
The result was measurable: when personality traits were incorporated into the Random Forest prediction model, prediction variability decreased by 36%, making the system significantly more stable—even though the average error increased slightly. In simple terms, the research statistically confirmed something experienced investors often say intuitively: founders with low neuroticism and high conscientiousness—calm, organized, and execution-focused—have higher chances of success.
The research starts from a basic assumption: startup viability is not determined by numbers alone. It is determined by the combination of three dimensions—market desirability (Desirability), technological and operational feasibility (Feasibility), and financial viability (Sustainability) itself.
The AI architecture developed combines four different data sources: business plans processed with natural language processing techniques, digital footprint from websites and social media, structured financial indicators, and—innovatively—founders’ personality data from MindEscape.
In experiments with seven machine learning algorithms on data from real Greek startups, ensemble methods—particularly the Random Forest algorithm—achieved the highest accuracy in predicting viability scores. The models identified that profitability, the existence of a minimum viable product (MVP), and satisfactory annual revenue above €57,000 are critical variables for startup longevity.
This academic research does not remain at the theoretical level. Dr. Takas conducted it with twenty years of active presence in the Greek VC/PE market as Managing Partner of THERMI Venture Capital Management—one of the country’s most stable investment firms, with roots in THERMI Business Incubator and three complete funds behind it.
Tools like MindEscape can function complementarily in the due diligence process, offering an additional, scientifically documented perspective in evaluating potential investments. As the dissertation concludes, AI does not replace the professional analyst’s human judgment, but provides valuable tools and data that reduce subjectivity and mitigate cognitive bias.
Read the full article on STARTUPPER.
© 2026 SeriousMind. All Rights Reserved. Privacy Policy