AXWAY column, by Emmanuel Méthivier, Business Program Director

Prospective: What will become of our banks if ChatGPT becomes my bank advisor?

In full digital transformation, the banking sector is moving gradually – and inexorably? – towards a “Bank as a Service” model. The ever more significant advent of generative and conversational AIs in customer relations suggests that tomorrow our bank advisor will be an AI. But then what about real “human” advisers?

During the last exhibitions dedicated to banking IT, all of the presentations we attended suggest that there is a triple subject of customer dissatisfaction with their relationship with the major Banking/Insurance groups.

First of all, the client no longer perceives his adviser as a trusted interlocutor but as a salesperson who tries to sell him as many products as possible on which he is commissioned, without taking into account (or too rarely) the client’s interest. Secondly, he reproaches him for not playing his role of intimate assistant in the important moments of life, which nevertheless represents a major expectation for a majority of individuals. Finally, digital self-services are not up to the standards imposed by mobile applications, forcing customers to go through the agency box. Patrice Bernard the very analysis as a “failed digitization”.

These criticisms, clearly exposed in a context of crisis of confidence, show the inability of financial services to correctly use data for the benefit of customer service – even when, protected by the GDPR, the customer consents to the use of his personal data. in return for a quality, personalized service with real added value.

Are the banks ready to disappear in order to survive?

This question is legitimate the evolutions are great in terms of generative AI and conversational. Indeed, it now seems certain that in the short or medium term, the various “Chats” – ChatGPT of course, but it is not the only one – will be much better advisers than the interlocutor available in the agency. Better trained, better informed about client habits, available, visionary, he may not be able to replace the wealth management expert, but he is quite capable of dethroning the incapable ‘ordinary’ advisor to understand their role in the customer relationship.

So yes, it is necessary to anticipate this new paradigm in which the customer relationship will be managed by an advisor boosted with AI. But under these conditions, how can a bank survive in this ecosystem?

One response may involve its ability to disappear, at least frontally, to accept disintermediation, direct interaction with the client, to dissolve into the daily life of its clients to retain control of historical financial services. Because a bank is a service provider, a function that takes on its importance in a more global moment of life.

Transactional and generative AI, a new generation of life advisor?

Banks had already taken advantage of the APIsation of the market and the Fintech ecosystem to implement new innovative services more quickly, such as payment terminals (Sumup, iZettle) and to refocus their role on certain more strategic activities. and for which they are more legitimate.

This acceptance of relegation as a service provider of a transactional, generative AI capable of taking on the role of life advisor, again requires a controlled and secure opening of its core system, thanks to APIs. This is necessarily accompanied by an additional layer, capable of opening up interaction with AI, namely marketplace tools that allow the banking institution to transform the technical API into a digital product. These DSM 1.0 (Digital Service Marketplace), showcase of digital services for Business Developers and human developers, will have to evolve, transform, to simplify their discovery (parsing) by AI and the consumption of their exposed services .

In Indeed, today, the maturity of the existing technology makes this transition possible and makes it possible to fully feed one or more AIs for a “perfect and complete” customer service: banking and financial products exist as “IT products”, links between AI et banks are possible, the security of exchanges and the confidentiality of data can be guaranteed, and current IT systems allow easy support for existing standards (DSP2, RTS, etc.) and the various banking connection networks ((SWIFTnet, SFTP, EBICS…).

But for AI to become a good “bank advisor”, banks will have to understand how AI will fetch information. Like a “chess player” AI who, once the rules of the game have been learned, will then focus on building her own strategies, a “banking advisor” AI will understand the issues, have access to customer information and then develop research strategies to offer the most suitable financial products. And the logic of an AI is not the same as that of a human. The bank will then have to create marketplaces of digital services designed specifically to match the logic developed by the AI.

In a ChatGPTized bank, what place for the human?

If the task of “bank adviser” is transferred to an AI, the question of the role of humans will be raised.

The very profession of banking adviser has had to and has known how to evolve over the last few decades. Until the early 2000s, we went to see our bank advisor in the branch to obtain an account statement or to withdraw money. These functions have gradually been transferred to ATMs. With the advent of AI “bank advisors”, the role of the human will once again evolve towards more added value.

Thus, with this in mind, 90% of the responses to questions and requests for advice from bank customers will come from AI. The human, on the other hand, will have to concentrate on second-level advice, in particular the wealth management or wealth management. Beyond the complexity inherent in these files, a psychological dimension will also come into play: a client ready to invest 1 or 2 million will wish to speak to a human being and not an AI. And even in this field, AIs will be able to provide research and analysis capacity to create a new class of human banking advisor: the augmented advisor, who will draw his advice from the information transmitted by AIs and from their expertise.

So how should banks evolve?

That’s the whole question! While sectors such as Travel and Tourism have already integrated DSM services into their business model, no bank has yet made the leap to DSM 2.0.

One thing is certain: there is a challenge of optimizing interactions that are very different from human interaction for this new generation of technology. An AI is a client for whom color, design, human accessibility are of little importance. It is necessary to understand its mechanisms of discovery, of use of the digital products which are placed at its disposal, via these DSM 2.0. On the other hand, it will be more difficult to subscribe to a paid plan… at least initially.

While the chapter on marketing studies relating to digital products is beginning to be written, the subject of the next chapter is already known: that of marketing to AIs, to identify their logic of research and use of data. Still beautiful things to discover.