ISFB Insight
Observing to Support: The ISFB Observatory’s Approach to AI in the Financial Sector
May 22, 2026
It has become necessary to assess the impact of artificial intelligence on the banking and financial sectors. But before diving into broad future scenarios, a simpler question must be asked: what are we really trying to observe?
AI is already transforming the workplace. We all see it in our daily lives. It speeds up certain tasks, changes work practices, boosts productivity, and forces the most forward-thinking organizations to rethink some of their processes. The pace of change is rapid—even breakneck. In banking and finance, where work is often highly standardized, this evolution warrants close attention. The question remains: what exactly are we talking about? Not AI itself, but what it changes as it is implemented. Tasks, skills, professions? Employment, competitiveness, or the skills each of us uses every day?
It is important here to distinguish between levels that are too often conflated. A task can be automated without an entire profession disappearing. A skill can become more important across the board, without being specific to any one profession. And two people in the same profession may be affected by AI differently, depending on their experience, their autonomy, and their relationship with the tools. This is why the impact of a technology on work cannot be measured as a simple variable. In the past, we contrasted routine tasks with non-routine ones; now we contrast substitution with complementarity. Yet a profession whose tasks are at risk can be augmented rather than replaced: the tool then takes over part of the work, freeing the expert to focus on what they do best.
The primary tool we use at the ISFB Observatory is the perception survey. We do not break down occupations into basic tasks to estimate exposure. We ask practitioners about what they experience, observe, and anticipate. This choice has a consequence that must be clearly stated. When a study reports that a certain percentage of jobs could disappear, caution is warranted: a perception survey does not predict what will happen; it reflects what respondents believe might happen. This data remains central—and indeed very valuable—since perceptions shape behavior, just as fears can either slow down or accelerate change. But it concerns perceptions, not the future of work.
This caution is all the more warranted given that the available data is often incomplete and of varying reliability. Each data collection method has its limitations. Macroeconomic indicators arrive after the fact, sometimes two or three years later—often too late to inform a decision at the moment it needs to be made. Questionnaires sent to a sample of employees gather perceptions on questions that are sometimes biased, and the most striking results too often present a simple correlation as a cause-and-effect relationship. Group workshops, which aim to collect data on a given issue in a semi-structured manner, offer access to concrete experiences or the interpretation of weak signals, but also carry their own biases: group effects, conformity, and fluctuating engagement. The point is not to reject any of these methods—which we use in our work—but to understand what each one allows us to consider reliable or not. An analysis never fully describes a situation; it describes a representation of it. A graph structures a line of thought and helps spark a discussion, but it can become problematic when it suggests it predicts the future with a precision that observed reality does not always allow.
When faced with a transformation that no one fully understands, the most effective response lies in fostering dialogue among stakeholders. This means giving a voice to experts on the ground, facilitating the exchange of perspectives, and helping to shape informed opinions where there were previously only isolated insights. After all, a system undergoing transformation does not rely solely on the accuracy of its predictions; it depends, above all, on its ability to share information and adapt collectively.
At this stage, the ISFB Observatory’s initial work therefore focuses on these preliminary steps: what are we talking about, what does the existing research say, and what biases should we account for. Rather than rushing to model a rapidly evolving phenomenon, we favor a pragmatic approach: literature reviews, perception surveys, analysis of simple indicators, interviews, feedback from the field, and cross-perspective analysis. Our goal is not to produce a model, but to contribute to the debate—modestly yet usefully—alongside other economic and academic stakeholders. A serious, ongoing conversation, grounded in the field, can guide decisions and ultimately support the allocation of resources toward developing the necessary adaptive capacities. That is the very role of the ISFB Observatory.
AI will transform the banking and financial sectors. But this transformation will not be limited to a single model. It will be systemic and evolving, characterized by feedback loops, unintended uses, resistance, and successive adjustments. The challenge is not to wait for a model to tell us what to do. It is about observing, understanding, and taking action in an organized manner. For individuals, this will involve training, experimenting, and developing new adaptive skills. For employers, it means supporting training initiatives by allocating substantial budgets to them. And for a community, it means establishing sector-specific forums like the ISFB Observatory where it can plan its own transformation.
Rather than predicting the impact of AI on jobs, the ISFB Observatory is taking the opposite approach: bringing together experts from the financial sector and facilitating their dialogue. For it is through this dialogue that a shared understanding is built, and it is from this understanding that the actions needed to support the necessary adaptation will emerge.
© Institut Supérieur de Formation Bancaire (ISFB). All rights reserved.
The analyses and content published by the ISFB may be quoted or reproduced in part, provided that the source is clearly mentioned. Any full or substantial reproduction of this article in another medium or format is subject to the prior written authorization of the ISFB. In order to facilitate reading and without any intention of discrimination, the masculine gender is generally used, in accordance with the grammatical rule that allows it to be used as a neutral value to refer to a group of people comprising both men and women. This publication is intended for ISFB members and their employees in Switzerland, as well as anyone interested in finance in Switzerland. It is not intended to be read or distributed in any jurisdiction where its distribution would be prohibited.
