Watch the session
Disruptors in Modern Finance with Omar Selim and Dr. Henning Stein.
Disruptors in Modern Finance
New York Technology Innovation hosts a session titled Disruptors in Modern Finance, featuring Omar Selim, CEO of Arabesque Group, in conversation with Dr. Henning Stein, Chief Innovation Officer at 1BusinessWorld.
The Trigger Behind Selim’s Focus
Selim frames his move from investment banking into building Arabesque around a conviction that technology can reshape an industry by changing both inputs and operating models. He points to the automotive industry as an example of rapid change driven by electric engines and self driving technology, then applies the analogy to finance. Asset management, in his view, is heading toward a comparable shift in how decisions are formed and executed.
The Three Disruptors That Shape the Thesis
Selim keeps the session organized around three forces that reinforce each other. Non financial data expands the information set used for investment decisions, he says. Artificial intelligence makes that expanded set usable at scale, he adds. Lifestyle change reflects the expectation that investors express personal preferences through portfolios without giving up performance, he says.
Sustainability Data as Decision Relevant Information
Selim describes research with universities aimed at understanding whether a company’s sustainability performance connects with market outcomes such as stock prices and credit spreads. He references a study with Oxford University’s Smith School that becomes known as From the Stockholder to the Stakeholder and describes its findings as mixed across sustainability objectives. He then explains why within industry comparison matters. A best in class approach ranks companies within the same industry rather than comparing sectors with different constraints, he says, and that method shows that companies attending to environmental topics, human rights, labor rights, and governance can outperform peers that do not.
The Mining Example That Connects Sustainability and Management Quality
Stein brings the discussion into mining and describes the difficulty of forming a deep understanding of what happens on the ground. Selim emphasizes correlation rather than causality and rejects a simple mapping between sustainability measures and stock prices. He offers a choice between two mining companies with similar financial outcomes over five years, where one legally disposes harmful byproducts into water and the other recycles them. Stein states that he favors global standards rather than firms that rely only on local legality. Selim argues that the recycler can signal a management team that uses technology to solve constraints rather than externalizing them, which can indicate more innovative and creative leadership.
Data Integrity and the Limits of ESG Scores
Selim argues that sustainability data only becomes useful when inputs are transparent and usable inside repeatable decision processes. He describes ESG information as often delivered in long report formats that resemble credit write ups and remain difficult to operationalize. He cites an MIT study he refers to as Aggregated Confusion and highlights limited alignment among ESG data providers, meaning vendors can disagree while presenting outputs with the confidence of traditional ratings. Selim argues that careful ingestion matters because apparent precision can conceal disagreement.
Building ESG Book Around Transparency
Selim describes founding a business within his group that evolves into ESG Book because he cannot find sustainability data that is transparent at the raw data level. He says the effort begins with aggregation to understand reliability and then expands into collecting proprietary sustainability data. He describes the organization as growing to about 230 people and building roughly 450 data points for up to 10,000 companies. He defines the data categories he focuses on, including business involvement and revenue sources, tenets of the UN Global Compact such as human rights and labor rights, environmental and corruption factors, and emissions and carbon data.
Best in Class and Exclusion as Separate Portfolio Logics
Selim separates best in class ESG analysis from business exclusion decisions. Best in class ranks companies within an industry, he says. Exclusion reflects which industries or revenue sources an investor wants exposure to, he adds. He illustrates the distinction with a contrast between a tobacco company that is exceptionally well managed across environmental and governance dimensions and a solar company that secures contracts through corruption, mistreats people, and performs poorly on recycling. Selim argues that one decision evaluates what a company produces and the other evaluates how it operates, and that investment clarity improves when those drivers remain separate.
AI as Capability Rather Than Branding
Stein challenges the market’s broad use of the term AI and notes that predictive algorithms and quantitative methods have existed for decades, including in insurance, without being labeled as AI. Selim agrees that AI is often used as marketing and distinguishes earlier approaches from current capability. A decade ago, he says, the industry has mathematics and strong talent but lacks the computational power and data capacity that make systems decision competitive. He argues that today’s shift comes from teaching systems to learn rather than specifying patterns to follow, allowing models to develop their own understanding of correlations and strategies when fed enough data.
Autonomous Investing Through AutoCIO
Selim argues that the defining question is whether a computer can take an investment decision. He notes that AI is widely deployed in back office and middle office functions such as fraud checks and risk management, yet he positions the central shift as decision making itself. He describes building Arabesque AI and a platform called AutoCIO as a holistic approach to emulating the investment process. He frames the industry’s trajectory as moving from active to passive and then toward autonomous investing. Autonomous investing, as he defines it, combines the quality of a human decision process with the rigor of a rules based data system at a fraction of the price. Selim says the system processes 32 billion data points every day and delivers equal or better performance than a human in about 81 percent of cases while staying true to the investment process, beginning with rebalancing improvements and extending toward new strategies over time.
Hyper Customization and Lifestyle Preferences
Selim argues that many people feel overcharged and underserved by finance and says the industry’s language remains opaque and complicated, stretching decision cycles. Passive investing grows as a response, he says, yet buying an ETF does not meet the demand for portfolios that reflect what people care about. Selim describes hyper customization as the direction, with AI managing portfolios on a phone and implementing preferences directly into the investment process. Stein tests the idea through his own example of veganism and preference based exclusions. Selim argues that AI enables customization without giving up performance when portfolio construction remains disciplined and outcome focused.
Institutional Alignment Without Sacrificing Performance
Selim extends the discussion to institutional investors and recalls conversations where investors argue they can only focus on shareholder value optimization and cannot consider climate and environmental factors. He argues that the direction shifts toward sustainability treated as a function of performance and alignment with values rather than marketing. He uses the example of a university endowment that fights cancer caused by smoking while its pension holdings include tobacco, arguing that better data clarifies revenue sources and production methods clearly enough to support implementation that matches mandate. Selim also argues that simple exclusion reduces optionality and can lead to underperformance during industry rotations, while AI enables investors to remove exposures and then identify compensating exposures that match risk profile and performance characteristics.
Regulating Autonomy in Finance
Stein raises regulation by referencing how sustainability frameworks can become difficult to navigate, then asks how Selim views regulation for AI. Selim argues that sustainability rules benefit from simplification and says AI regulation is difficult to design because the technology evolves quickly. He rejects a world where AI runs entirely without guardrails and uses self driving cars as an analogy for how trust forms around autonomy. Statistics can suggest that autonomy reduces accidents overall, he says, yet failures still occur and individual incidents can reshape confidence. Selim argues that finance moves toward digital asset management where a human portfolio manager no longer presses a button to approve each decision, and he frames the regulatory task as designing a pathway from the current model to the next model while weighing broad benefits against inevitable costs.
Integrated Investing as the New Baseline
Selim and Stein bring the session together around integration. Selim argues that non financial data becomes decision usable when inputs are transparent and noise is reduced, and that advanced AI reinforces investing by emulating human decision processes and enabling portfolios tailored to preferences while maintaining performance discipline. Stein reinforces the connection between high quality data, sustainability considerations, and advanced AI as the foundations of modern portfolio thinking. Selim anchors the outlook in an industry trajectory that moves from active to passive and then toward autonomous investing, with AI serving as the decision engine that translates expanded information into more precise portfolio action.
Sources
Continue on the session page
Access the full session page for additional context.
