Before the financial crisis, I spent years meeting customers across the capital markets while working at a major credit rating agency, Moody’s. At that time, I believed there was generally only one way to evaluate credit because that was the world I knew. But around 2007, I started meeting institutional investors who viewed the world through a different lens. They were identifying risks others missed and making calls that would define their careers and would deliver their investors unprecedented returns. This was the first time I realized there were entirely new ways to understand credit.
Around that time a colleague recommended a book called Black Box Thinking, which tells the story of Abraham Wald during World War II. Engineers wanted to reinforce the bullet-ridden sections of returning planes, but Wald showed that the real danger was in the untouched areas because those were the planes that never returned. It was my first real lesson in survivorship bias. In credit, we often focus on what is visible while missing what quietly fails.
During those years I began asking quantitative friends, in what felt like my Big Short moment, whether it was possible to build a system far more powerful than the rules based legacy qualitative and quantitative models that defined the industry. Most said the same thing. Possible, but not at the level of improvement that would change the world.
Everything shifted when I later worked alongside Google Kaggle Grandmasters at a startup that eventually became an AI unicorn, DataRobot, (think of Palantir but with a major part of their business in financial services). What this exceptional team built in the early years of AI was extraordinary. The pace of innovation from day to day, week to week and month after month was unlike anything I had experienced working at a Fortune 500.
It was in that environment that I met colleagues who would later become my co-founders. Joe Burdis is an AI expert with a Ph.D. in math and CFA charterholder that worked at Deloitte and SAS, where he built and deployed mission-critical analytical systems used by major financial institutions. Arbi Abeshi came from Morgan Stanley and Goldman Sachs before joining the AI unicorn world, where he delivered large scale, high stakes AI engineering solutions in financial services.
Their ability to build AI systems at scale, architecting, engineering, and delivering technology that consistently solved problems others considered impossible, was extraordinary. They were absolute legends, and witnessing what they could deliver fundamentally changed my understanding of what elite engineering and applied AI could achieve.
It was in that environment that the idea of an AI powered credit system resurfaced, but the technology was not fully ready.
Then GenAI arrived. Suddenly the system we had imagined years earlier was not only possible but within reach. In July 2024, we founded AIR Platforms to solve what we believe is the ultimate math problem in financial markets.
The world is becoming more complex. Macro conditions shift faster. Information is noisier. AI has moved from experimentation to production. As this happens, the magnitude of both success and failure will grow. The next wave of credit dislocation could reach companies, institutions, and many of the people who rely on them that could have existential consequences. But that outcome is not inevitable.
Think about transportation. Over the last century, countless lives were lost because of missing safety systems and incomplete information. Now imagine if a company like Tesla had existed one hundred years ago. How many lives would have been saved if vehicles learned from every close call and made every driver and pedestrian safer. Credit is in that moment right now.
If automobiles and even spacecraft can navigate themselves using continuously learning AI, then credit ratings should be able to do the same.
Credit determines how capital flows. It decides which companies grow and which fall behind. Yet for more than one hundred years, credit ratings have relied on static methodologies and slow, manual refresh cycles. Bias seeps into frameworks. Early warning signals are often missed. Risk becomes visible only after it is too late.
But markets move in real time. Risk moves in real time. Institutions need insight that moves the same way.
AIR evaluates every company every single day across public and private credit. It learns, recalculates, and signals early. It analyzes without bias. It operates continuously. When the credit cycle turns, AIR does not blink. It functions like a modern AI powered Iron Man credit suit that gives analysts superhuman awareness, not by replacing them but by amplifying their abilities.
This is the foundation behind our announcement today. AIR has raised $6.1M in seed funding, co-led by Work Bench Ventures and Lerer Hippeau, to modernize how the world understands credit. We are ecstatic to partner with both firms!
AIR has already shown what is possible.
When First Brands, a large auto parts manufacturer, began deteriorating toward default, one of our customers was immediately informed. AIR assigned a rating aligned with the lower end of the speculative grade spectrum, comparable to CCC, well before the broader market recognized the stress. These signals proved accurate and revealed risks that legacy approaches missed.
We are now partnering with institutions managing billions in assets. This includes CLO managers, BDCs, banks, pension funds, and Fortune 500 enterprises. Customers use AIR for daily ratings, early warning, sensitivity analysis, portfolio triage, compliance and for a variety and multitude of use cases.
We built AIR because we have lived every part of this journey from the depths of the financial crisis, watching institutional investors spot what the market missed, from learning Abraham Wald’s lesson about the planes that never returned, and from building world-class AI systems with Kaggle Grandmasters that solved problems once thought impossible. Those experiences taught us that credit needs a new way to see risk, the same way Tony Stark needed the HUD to see what the human eye could not.
What we have built to date would not have been possible without the conviction and support of our early investors, including Work–Bench, Lerer Hippeau, and a group of angel backers who understand both credit and enterprise AI at a deep level. Their belief in this mission from the start gave us the ability to build AIR with the rigor, scale, and focus the problem demands.
If you are tired of noise and want partners who understand credit at a deep level, who have actually delivered AI in the most complex financial and variety of other environments, and who see their role as trusted advisors rather than vendors, we would be honored to partner with you. No team is more focused, more experienced, or more determined.
Credit is the ultimate math problem and AIR was built to solve it continuously, objectively, transparently, and without bias for a world that can no longer afford blind spots.
LETS GO!!!
