AI in Wealth Management: How ChatGPT and Anthropic are shaking up the traditional banking relationship
Friedhelm Schmitt
In the debate around Artificial Intelligence in Wealth Management, the focus is almost always on tools, assistance systems, automation and efficiency. Friedhelm Schmitt, co-founder and CEO of fincite, believes this is the wrong perspective. The real question is not which tools banks use, but what their business is actually based on, and this advantage has already shifted. With fincite • cios, fincite develops a modular wealth management software that is used by over 9,000 wealth managers in Europe and is part of the Harvest Group.
Banks' knowledge edge is no longer exclusive
For decades, a bank's business model was based on a knowledge advantage. The bank understood markets, products and financial contexts better than its clients, and this asymmetry created the value of advisory services. When providers like OpenAI integrate financial functions directly into ChatGPT today, or Anthropic brings its model into corporate finance departments, the logical conclusion is that the traditional banking relationship is beginning to totter.
However, software inserting itself between provider and customer is not a new phenomenon. We experienced it in the travel business with Booking, in retail with Amazon and in banking itself with comparison portals like C24.
What is remarkable is something else: for the first time, a system is available that can capture an individual's financial situation more comprehensively than a bank does today. Displacement by an intermediary is therefore not the actual threat. The decisive factor is that the understanding of the client is no longer exclusive to the bank.
Why categories like risk classes no longer reflect clients holistically
The way banks capture their clients today has remained remarkably coarse-grained. A client is categorised by age, risk class and investment horizon, and their advice is derived from this classification. They are then deemed risk-tolerant or risk-averse, as if this were a permanent characteristic and not a snapshot in time. This has little to do with a person's reality, and even less with how modern systems function.
An Artificial Intelligence can holistically capture what phase of life a person is in, how their situation is changing and when they are about to make a decision against their own interests. In doing so, it already provides a more differentiated assessment today than the traditional advisory process in most banks offers.
Therefore, those wealth managers who view their role as a platform will prevail. They accompany the client across their entire wealth, intervene situationally and build trust over long periods, instead of pigeonholing the client into an existing grid once a year.
The three layers of the banking stack in the AI era
Anyone who thinks the technological structure of the AI era through to its conclusion will recognise three layers building on top of each other. For banks, it is not crucial whether these layers exist, but which of them they control themselves.
The data aggregation layer forms the foundation, because language models and business applications only unfold their potential once they can access reliable client data.
The business application and AI layer generates the actual utility, either through the bank itself or via partnerships, and is complemented by the models above it.
The regulatory and compliance layer must absolutely be controlled by the banks themselves, as it is the only one that ensures everything above it remains trustworthy, auditable, traceable and regulatorily sound.
This final point in particular is underestimated in current discussions, and only a few banks have it on their strategic radar. So far, the industry has mainly occupied itself with preparing and displaying information. However, the next development step of AI lies in action: in advisory systems that execute recommendations independently, and subsequently in the active management of portfolios. Very few institutions are prepared for this development so far.
Wealth Aggregation as a strategic factor
Europe has missed a strategic opportunity with Open Banking. In both regulatory implementation and political lobbying, the opening of financial data was slowed down rather than driven forward. However, developments in the United States clearly show where the journey is heading: the aggregation of financial data will happen, and the only open question is which player will facilitate it.
As long as banks or regulated aggregators control this layer, they retain sovereignty over access, the granting of permissions, governance and customer consent. This means they keep actual control over who can use which data and for what purpose.
For this reason, bank-operated wealth aggregation is the first concrete step. It is the gateway to the AI layer and falls under the bank's responsibility. Those who surrender it risk becoming pure infrastructure in the long term, whilst others shape the customer relationship. This is precisely the layer around which fincite • cios is built – a modular wealth management software that operates API-first and can be integrated into existing core banking systems.
Memory as the next stage of development in advisory services
The most strategically far-reaching aspect concerns the ability to store context over time. I consider this form of memory to be the next major development stage of Artificial Intelligence. Human identity is largely created by the fact that experiences shape us, that we connect them over the years and derive patterns from them. Today's systems do not yet fully possess this capability, but the development is clearly moving in this direction.
Precisely at this point, the weakness of banks becomes particularly evident. The picture a bank has of its client is usually limited to income, risk class, marital status and a few meeting notes. It remains a static snapshot that corresponds neither to a person's reality nor to the logic by which future AI systems will operate.
What banks must therefore build is a layer that captures a client's context over their entire lifecycle, from major life events to changing market phases and shifts in behavior. Only on this basis does highly individualised advice become possible.
Conclusion: What Artificial Intelligence in Wealth Management means for European banks
Artificial Intelligence in Wealth Management has reversed the information asymmetry that banks lived off for decades. The advantage is no longer automatically with the bank, but with the system that understands the client best. Those who want to win back this advantage must keep data aggregation, the compliance layer and the memory of the client under their own responsibility. Wealth aggregation is the first concrete step towards this, and the technology needed for it is ready. The decision about this lies with the banks themselves.
