"Regulation occupies an extremely important place when integrating and using AI in the provision of investment services"

Cecabank joins other projects such as Inversis, Santander, BBVA... to personalise financial advice with AI. Where is the industry going with the use of AI?

GPT: Since 2023, when the emergence of Generative AI became evident, all industries have been adapting to how they can leverage it. The financial sector, which is highly information- and content-intensive, is set to be one of the biggest beneficiaries of it. I believe we are currently experiencing a dual process: on the one hand, to have a vision of how far this technology will transform or revolutionise the long-term financial service; and also to get to work and start making projects to apply this technology to different use cases.

Cecabank: There are two final objectives. On the one hand, we want to harness GAI to increase the capabilities and benefits we offer our customers through our services and products. We also seek to identify and address improvement routes in terms of process efficiency, incorporating tools based on Artificial Intelligence that are suitable for each purpose.

In addition to the above, our exploration pursues other intermediate objectives, such as obtaining in-depth knowledge and understanding the characteristics and nature of this technology, studying the best practices to carry out a smart and prudent integration of GAI within our organisation, systems and processes. We already know that this technology has enormous utility and potential, but the challenge now lies in being able to apply it effectively in our business.

GPT: In the first phase, GAI will be highly effective simply by centralising all the information and content of a financial institution, such as its catalog of products and services. And do this by complementing it with training content for the bank's employees. Long before a client interacts directly with a bank’s AI assistant, employees of these institutions will be the first to use these tools. This will allow them to access better information and training about the institution’s services, enabling them to personalise offerings more effectively to meet each client’s needs.
In the next phase, clients will be able to interact directly with AI agents from each institution. These agents will not only provide information but also offer advice in accordance with the institution’s rules and compliance regulations.

Cecabank: By its nature, Generative AI provides special value in all processes that work with inputs and/or outputs based on unstructured data. These normally cannot be handled by deterministic software programs and are solved using inherently human skills, such as language and reasoning ability. Initially, these unstructured data used to be considered in text format. However, possibilities grow exponentially if we consider trends towards multimodal Artificial Intelligence models, able to work simultaneously with text, images, audio, video, etc.

From the above, it follows that applications in processes, in an efficient way, are infinite: Analysis and response to customer queries while maintaining the personalised treatment they deserve, analysis of product information from non-standardised documents, support in project execution, integrating AI as another team member...

On the subject of portfolio management, do you think you can teach the machine how to manage uncertainty in the markets and global society?

GPT: The main revolution of Generative AI does not, at least for now, relate to market prediction, portfolio optimisation, or similar areas—that will come from other types of AI. Generative AI hype is so huge right now that many opportunists are claiming it can handle those tasks well too. What GAI brings is a revolution in language and communication. And what this could lead to (as we have already seen in the crypto world) is an AI agent that learns and persuades us that it has those capabilities—ultimately profiting from those who naively believe it.

Where does regulation stand? 

Cecabank: Regulation occupies an extremely important place when integrating and using AI in the provision of investment services. Within the regulatory spectrum, we must consider two different categories: On the one hand, the regulations that specifically and expressly regulate the use of AI and, on the other hand, the rest of the rules that could be indirectly impacted. . And in fact, in the CNMV Sandbox, we are seeing very interesting cases.

On the subject of specific regulations, the European Parliament approved the Artificial Intelligence Act (AI Act) in March of this year. This law seeks to create a harmonised legal framework to guarantee the protection of fundamental and public rights. It defines and categorises AI systems and models through a risk-based approach, understanding the risk from the point of view of harming citizens' rights. It prohibits unacceptable risk systems and sets, according to their risk level, the requirements to be met by the rest of the AI systems. Although it is a law that seeks to promote innovation, it also establishes requirements on transparency and data governance and on the needs of supervision and compliance.

With regard to other regulations that are not specific to AI, it should be noted that it is necessary to analyse how an organisation, systems and processes of an entity are modified when integrating AI functionalities. The aim is to identify which of these changes entails the need for regulatory adequacy to ensure regulatory compliance, e.g. with regard to personal data protection.

How does this benefit the end user and what type of user will benefit the most? 

Cecabank: The benefits of GAI for users will depend heavily on the type of use and application in question. However, in general terms, GAI will make it possible to:

• Specifically increase the level of personalisation of services, in all the end customer segments
• Provide more and better information based on the same data available to the bank, adapting the responses to each of the user's queries and information requirements (e.g. Through chat tools, configured with very specific data and objectives)
• Gain time and reduce efforts in the analysis of information
• Facilitate decision-making
• Improve employee and customer satisfaction levels, eliminating certain processes that drive failure

In potential terms, all users (managers and other employees, customers, suppliers, etc.) could benefit to the same extent from the improvements planned with the application of GAI, provided that specific problems of each type of user are addressed.

Digital assets, tokenisation... Where will the next steps in the sector go? 

Cecabank and GPT:  In our opinion, digital assets are already a reality and have come to stay. Tokenisation of assets is a resource that can give capital markets unprecedented versatility. In the field of digital assets and tokenisation, we see a frantic activity in different directions, with a long projection in each of them:

• Creation of new organised markets based on DLT, under the pilot scheme
• New asset custody services
• Introduction of new figures in the value chain (e.g. ERIR)
• New investment products (investment funds that invest in cash cryptocurrencies)
• Introduction of CBDCs (Central Bank Digital Coins-central bank digital money)
• Tokenisation services of financial instruments (fixed-income or equities) for issuers without access to the traditional capital market (small and medium-sized enterprises)
• Introduction of digital assets in the catalogue of services of consolidated and generalist credit institutions, etc.

What we are seeing is that the market, through its different players, is making progress in parallel in all the aforementioned lines, seeking to create and/or consolidate new ecosystems that provide a differential value on the current market structures, while at the same time seeking to strengthen these new systems to strengthen these systems to provide them with the confidence they require to attract a broad audience.