An Interview with Holger Bartel, RealRate CEO: ‘We use Explainable Artificial Intelligence to Create Fair, Unbiased, and Ethical Company Ratings’

“The structure imposed by our expert model enables a holistic valuation and causal explanation.”

RealRate, a rating agency using Explainable Artificial Intelligence (XAI), is one of the most promising German AI startups. It was incorporated in 2016 and is headquartered in Berlin, Germany.

Holger Bartel, RealRate CEO, spoke exclusively to The Silicon Review.

Q. Please give us a brief overview of your education and professional career.

I studied economics and was a student teacher at Humboldt University, Berlin. Later, I finished my Ph.D. thesis in statistics in 1999.

While pursuing doctoral studies, I founded my first startup Prozentor in 1997, a fintech providing stock market forecasts. We raised 1M Deutsche Marks in 2001, won Deutsche Bank and Sal. Oppenheim as customers. We also created capital market products. All this was when the term ‘Fintech’ or ‘Robo Adviser’ wasn’t even born yet.

Later, I worked in the insurance industry; the German Insurance Association and two other large German insurance groups. I was the head of life insurance mathematics, reporting to the board. I was also announced an appointed actuary for the groups’ three life insurers, and the property-casualty insurer.

In 2016, I joined the startup scene again, founding RealRate, the first rating agency based on Artificial Intelligence (AI). After three years of developing the AI software, we went live in late 2019, immediately winning our first customers.

In 2019, I co-founded yet another startup, taking again the role of CEO—DigiOptions—a blockchain fintech, providing options and betting markets on the blockchain. We won a grant from Investitionsbank Berlin. In 2020, DigiOptions was a finalist in the ‘Future of Blockchain Competition’ and was part of the EnCode Thundercore program. DigiOptions is accelerated by Startupbootcamp Netherlands and GSD Venture Labs from the US.

Q. Your focus has been on risk management and financial market analysis. Give us anecdotes as to how this journey has been so far?

It was exciting with my first startup to provide financial forecasts. And our technology performed exceptionally well. However, when the dot-com bubble began to burst, our algorithms recommended selling everything. At the time, we lost many customers, many of them later wishing they had followed that recommendation.

Regardless, we stuck to our original approach: using modern technology (data science, blockchain, AI) to provide new, fast, unbiased, and fully automated services.

Eventually, that paid off. For instance, RealRate, providing ratings based on AI, immediately succeeded in the market.

Q. How did you find RealRate, and what was your thought process behind starting the company?

When I was working as an appointed actuary in a big insurance company, I was responsible for talking to the ratings agencies. One day our big boss came in and told the rating agency: “Listen, I am expecting you to give me a triple-A (AAA) rating, otherwise there are plenty of other rating agencies out there.” At that moment I figured that the rating won’t be fair but biased. That said, we at RealRate are committed to change that.

Besides, I learned that the traditional interactive rating process was highly cumbersome, expensive, and time-consuming. The agencies were under pressure to produce a good rating to be commissioned again next year. Therefore, they weren’t independent.

I thought about a simpler and unbiased approach of using AI to be independent, rating whole industries, and afterwards addressing the best-rated companies, offering them to purchase a RealRate seal of approval to strengthen their brand. Furthermore, we were the first who made Artificial Intelligence explainable and applied to ratings.

As a matter of fact, I was my first customer: As an appointed actuary at an insurance group, I needed to perform a benchmark analysis of the group’s financial strength. I made the AI explainable for the first time; I understood the company’s strengths and weaknesses compared to the market average.

Fair, interpretable ratings and no black box marked the birth of RealRate.

Q. RealRate combines expert knowledge and Artificial Intelligence to create fair ratings. Please elaborate on this.

AI is transforming the world. From self-driving cars to making credit decisions in the banking industry, the technology proved its worth. But there always remains the question of “Why?” or “Is it fair?” We need to ask ourselves whether we can trust the technology. If we use the technology, we should be able to interpret why a certain decision was taken!

That said, RealRate offers explainable rating decisions powered by AI systems. The valuation logic (the RealRate model) is implemented to enable causal reasoning. Our software then generates an interpretation and a graph showing the real causes and effects of a company’s financial strength. The graph is colored, immediately showing the company’s strengths and weaknesses.

Astonishingly, you’re able to interpret a company’s financial strength within two minutes instead of reading 200 pages of the business report. This self-developed technology laid the foundation for RealRate’s unique positions and why it is standing out among its competitors.

Q. Your approach incorporates a holistic business model of the company instead of a simple weighting of key figures. How do you achieve this?

Classical ratings often use a bunch of key ratios to derive a rating. Instead, we use a whole valuation model. Starting from the basic input (certified public balance sheet data), we use our model to step by step compute economic figures, which eventually results in the company’s financial strength. The structure imposed by our expert model enables a holistic valuation and causal explanation.

Q. In the wake of the recent economic instability the world has been facing, what are your goals for RealRate? And what are your expansion plans for the company?

First of all, we hope that RealRate will add to increase economic stability.

Rating agencies play an important role. The financial crisis of 2008 was driven by overly optimistic assessments of the big rating agencies. Our technology will avoid any conflict of interest, improving information, and understanding, thus stabilizing the economy.

We are now entering the US market. It is a huge market with many listed companies and highly standardized electronic annual business reports, a perfect ground for our explainable AI rating system. We will cover dozens of industries and thousands of companies.

Q. What are the challenges you face in the industry and the solutions you offer to overcome them?

Regulation is a crucial aspect of the financial industry and ratings markets. Companies must further professionalize their risk management, strengthen their resilience, and build up trust. And banks, for example, will be obliged to explain, why they rejected a loan to certain individuals.

RealRate can add a lot to overcome these challenges, offering independent and interpretable ratings for companies. Companies could use this independent statement to convince customers of their excellent financial position and thus strengthen their brand.

Q. What are your projections for the insurance sector in the near future? Do you see more players entering the market?

We started by rating insurance companies. They heavily depend on capital markets, even though their original approach is to offer biometric insurance. Insurers are among the biggest capital accumulators worldwide. Their investments are crucial for the economy, but they require returns, especially if they provide financial guarantees like life and health insurers do.

Unfortunately, in the current economic situation, interest rates have been pushed down by central banks worldwide. This might help the banking industry but will poison the insurance industry in the long run.

I, therefore, expect that European pension and life insurers will leave the market. Probably this will be mainly realized by mergers and acquisitions.

Insurtechs, instead, will play an increasingly important role.

Q. Tell us more about the recent milestones you achieved.

RealRate’s innovative approach has been widely acknowledged:

• In March 2020, RealRate was elected one of the most promising German AI startups by the Applied AI Initiative.
  
  
• In October 2020, the RealRate AI software was the winner of the PyTorch AI Hackathon. More than 2500 teams submitted their projects.
  
  
• In October 2020: RealRate is announced “Most Innovative AI Financial Ratings Provider” by EU Business News.
  
  
• In November 2020: Startupbootcamp Fintech Amsterdam, the first publicly-listed accelerator globally, selected RealRate for its program. We also joined the GSD Venture Labs accelerator.

Q. What makes you different from traditional rating agencies?

In contrast to the big rating agencies, we have no human analysts. Instead, we use our award-winning software making AI explainable. There is no black box with RealRate.            

AI ratings are an exciting use-case for smart AI in finance. We will go fast to conquer that market.

“In contrast to the big rating agencies, we have no human analysts. Instead, we use our award-winning software making AI explainable. There is no black box with RealRate.”