The upcoming challenge in fintech is gaining insight into the individual behind the data.

      **TL;DR** Fintech has made financial data accessible but not comprehensible. Amr Mohamed, founder of Neumetria, contends that the next step should be to interpret the reasons behind transactions, rather than merely classifying them. As AI agents interact with financial systems, the depth of contextual understanding will influence the appropriateness of decisions. The article references the Federal Reserve (73% of adults feeling financially secure), BIS's research on AI in finance, and NIST’s AI Risk Management Framework.

      Fintech has spent nearly twenty years enhancing the accessibility, portability, and usability of financial information. Innovations like open banking, digital wallets, and connected financial platforms have created a setting where consumers can seamlessly transition among products, allowing authorized providers to access increasingly detailed information regarding their financial activities. According to the Consumer Financial Protection Bureau, over 100 million consumers have utilized authorized data access, granting third parties insight into their financial data.

      As financial institutions have witnessed a rise in available information, the complexity of consumers’ financial situations has also increased. The Federal Reserve’s 2024 Survey of Household Economics and Decisionmaking indicated that 73% of adults reported being financially stable or comfortable, with 63% stating they could manage a hypothetical $400 emergency expense with cash or its equivalent. These statistics reflect financial states at a specific moment, whereas household income, spending, and financial resilience are subject to continuous change.

      This distinction is significant, as a transaction is merely a data point rather than a complete narrative. A purchase reveals details like where money went, when it was spent, and the amount used. However, the implications of the same transaction can differ for individuals with varying income patterns, responsibilities, savings, risk tolerance, or financial objectives. Research from the Bank for International Settlements has emphasized the growing significance of data quality, governance, and management as AI becomes more integrated into financial services.

      Thus, the challenge for fintech is to evolve from simply collecting information to truly comprehending it. Amr Mohamed, CEO of Neumetria, posits that this represents the next critical layer of financial technology. He argues that, although the sector has developed extensive infrastructure for accessing and analyzing financial data, a notable gap persists between observing transactions and grasping the circumstances of the individuals behind them.

      Mohamed's insights stem from years of experience in creating financial products and infrastructure. His dual perspective as a fintech user and creator has underscored a recurring disconnect between access and context. “As both a consumer and a builder within fintech, I have noticed how quickly the industry has broadened access to financial products while the surrounding context has remained disjointed,” Mohamed states. “Consumers may engage with banking, investment, and digital asset products separately, but their financial choices are essentially interconnected. The consumer is navigating one financial life, regardless of the quantity of products involved.”

      This leads Mohamed to a concept he terms ‘behavioral understanding.’ He believes that the next phase of fintech will necessitate systems capable of unifying banking, spending, saving, borrowing, investing, and other financial activities into a continuously evolving overview of an individual’s financial status. The goal is to develop what he refers to as a dynamic financial state that adapts as behaviors shift.

      In Mohamed’s view, this entails moving beyond mere transaction categorization. Most systems can identify whether a transaction is for dining, travel, groceries, or another category. He asserts that the more crucial question is why the transaction took place and what it signifies within the individual’s broader financial journey.

      “Categorizing a transaction tells us what took place,” Mohamed explains. “The more valuable inquiry is what the transaction indicates about the individual’s circumstances, goals, and financial conduct. Context enables a financial system to progress from recording activities to understanding them.”

      This contextual layer could facilitate financial products and agents in assessing indicators such as income stability, liquidity horizons, spending confidence, financial resilience, and readiness for borrowing or investing. Two individuals may seem similar based on historical transaction patterns but possess vastly different current capabilities as one might have a stable income and a rebuilding savings buffer while the other faces declining income and mounting obligations. Mohamed argues that financial infrastructure should be adept at differentiating between these evolving situations.

      This importance amplifies as AI transitions from merely generating recommendations to directly interacting with financial systems and potentially acting through agents. Mohamed insists that while an AI system can process information, its reasoning is only as effective as the context it is given. Before an agent suggests saving more, transferring money, taking credit, or investing, it must have a trustworthy understanding of whether those actions align with the individual’s present situation.

      NIST’s AI Risk Management Framework stresses that trustworthy AI must prioritize reliability, transparency, explainability, privacy, and fairness in its operational context.

      For Mohamed, this underscores that financial understanding must be viewed as an infrastructure issue for the forthcoming era of agentic finance. This contextual layer could enhance various financial experiences, enabling products to evaluate if a customer is suitable for specific offers,