A few weeks ago, a colleague of mine on the board of CFA Society Spain told me that she had created ten agents based on ChatGPT. These agents carry out fundamental analysis based on the quarterly earnings releases of the listed companies within her investment universe, as well as drafting and presenting certain reports tailored to the needs of the firm she works for.
The pace of advancement in the capabilities of these systems is so astonishing that one barely has time to pause and reflect on the threats and opportunities they present. If one wants to remain relevant in this environment, the main recommendation from experts in this field is to devote at least an hour a day to learning about and using the most advanced models, approaching this with a mindset geared towards overcoming the continuous obsolescence we are likely to face.
The question for investment professionals has shifted from whether or not to incorporate agentic AI to which organisational model to adopt, given its growing ability to plan and execute increasingly complex tasks with a rising degree of autonomy. In other words, the change lies in the fact that, whereas in the past technology was used as a passive tool, it is now becoming an active collaborator capable of replacing tasks typically carried out by a junior analyst.
But when we speak about agentic AI, what exactly do we mean? This refers to a system that brings together a range of capabilities, such as the use of a large language model (LLM), access to external systems or databases, and the ability to plan tasks using evaluation and feedback mechanisms within a structured memory environment.
The savings in time and direct costs,and therefore the resulting increase in productivity, are considerable. It is true that certain roles may be replaced, particularly during the first five years of a professional career, but this will also lead to the creation of new roles within organisations. At the same time, for many employees, responsibilities and functions are likely to expand.
When an AI agent is capable of performing macroeconomic analysis, valuing companies according to specific criteria, integrating alternative data and news into its assessments, and identifying risks of different kinds, it allows the investment manager to focus on strategic interpretation and on reflecting on the suitability of an investment. Judgement and decision-making do not disappear.
In addition, the use of AI in this way can enhance the oversight capabilities of risk analysts, streamline compliance and reporting processes for regulators and clients, and significantly reduce administrative burdens, operational errors and turnaround times.
At the same time, it will be critical to implement robust human oversight and control mechanisms, particularly given the regulatory intensity of the financial markets industry, where technological governance will be indispensable.
In any case, ultimate responsibility for investment decisions and fiduciary duties towards clients will continue to rest with the investment manager. While their day-to-day work may undergo a more or less radical transformation, their core mission will remain entirely unchanged.
Recent history in financial digitalisation shows that the advantage does not usually lie in being the first to adopt new technology, but in integrating it coherently within the existing organisation.
Agentic AI represents a qualitative shift in how financial institutions can structure their work. It is not about replacing professionals, but about expanding their operational capacity through systems capable of planning and executing complex tasks.