Insights from Co-one’s Co-Founder on From AI Assumptions to Real World Readiness and Navigating the AI Turbulence w/ Mert Menekşe
Insights from Co-one’s Co-Founder on From AI Assumptions to Real World Readiness w/ Mert Menekşe
Mert’s path began at Istanbul Technical University, where he and his co-founder Arman won international aerospace competitions organized by NASA and ESA. After spending 7 years on the investor side at Vestel Ventures, Mert decided to cross to the other side of the table to become a better investor through hands-on experience. Co-one initially launched as a data labeling platform but quickly pivoted as the AI landscape evolved at breakneck speed.
In an era where AI models are updated daily but the risks of "hallucinations" and errors remain high, Mert transformed Co-one into an AI Testing and Analysis Platform. Today, Turkey's leading banks (İş Bankası, Akbank, Garanti) and e-commerce giants (Trendyol, Hepsiburada) use Co-one to audit their AI models for accuracy, error detection, and cost efficiency. Mert describes this process as "stress testing models to ensure they behave deterministically."
Having closed a strategic investment round during a period of capital scarcity, Co-one shifted its growth focus toward a global structure based in Estonia, leveraging the power of international networks. Mert noted that while Turkish corporations are "early adopters" of AI technology, local startups need more patience and a stronger global diaspora to achieve worldwide scale.
Reflecting on personal discipline and mental health, Mert shared that he and his co-founder manage stress by working out 5 days a week, integrating entrepreneurship into every aspect of their lives.
"We fill the gap between the assumption that AI models can do everything and their actual readiness for the real world through testing and validation, " Mert said, offering this message to fellow entrepreneurs:
"Don’t just follow trends; find your 'Ikigai' at the intersection of your talents, your passion, and what the market is willing to pay for. The AI world is going through massive turbulence; to survive and create value, you must ensure your models are truly ready for real world deployment."