The Cercle K2 does not intend to give either approval or disapproval to the opinions expressed in the p ublications (written and video) which remain the property of their author.
----------
The history of today's universal banks is made up of mergers of various entities, brands and international networks initially developed by business line with a view to customer proximity, sector or regional expertise, always subject to the requirements of local regulators.
In this structurally fragmented environment, data sharing between the various business lines of the bank as well as with customers and partners is emerging as a key issue for banks:
In this structurally fragmented environment, data sharing between the various business lines of the bank as well as with customers and partners is emerging as a key issue for banks:
- In the short term, the challenge is the development of consulting and artificial intelligence as a factor of agility and competitiveness for the bank in an environment that weighs on the banks' balance sheets.
- In the long term, the challenge is the positioning of banks in the data economy, the provision of trusted data in the banking sector and the possible diversification of banks in the exchange of data between customers from various industries.
The data-driven bank or the generalization of a banking advice centered on artificial intelligence as a competitive factor
In the short term, universal banks are hoping to leverage data to differentiate themselves through advice specific to each of their activities.
The rise in key rates and the resumption of inflation in a context of questioning the major geopolitical balances and growing risk aversion of investors in the face of hazards (particularly ESG) are weighing heavily on banks' balance sheets. The risk of bonds is increasing, while the performance of so-called risky products is burdened by new regulatory and market requirements.
The rise in key rates and the resumption of inflation in a context of questioning the major geopolitical balances and growing risk aversion of investors in the face of hazards (particularly ESG) are weighing heavily on banks' balance sheets. The risk of bonds is increasing, while the performance of so-called risky products is burdened by new regulatory and market requirements.
Encouraged by the latest European directives, new competition is developing in the area of payment methods, and banks must open up access to the details of their clients' accounts to specialized fintechs, which are likely to take over their now digital client relationships.
Of the four main banking roles: the lessor who provides funds on credit or in equity, the safekeeper who ensures the maintenance and securing of accounts, the diligence who ensures the liquidity and availability of funds where needed, and the advisor who guides your investments or allows you to liquidate them at the best price, it is from the advisor that most of the players hope for their salvation.
Indeed, advice is present in every banking business. Its revenues are based on commissions that follow inflation. Banking advice has a direct influence on the quality of arbitrage and client portfolios. Finally, it contributes to customer loyalty in an increasingly digitalized world where there is a risk of disintermediation.
It is about understanding behavior, analyzing and comparing performance, and evaluating the return and risks of a particular, even individualized, situation and context.
Now able to identify new customer behaviors and new customer segments, to anticipate long-term value and to provide relevant recommendations, Artificial Intelligence is the main lever for efficient and widespread banking advice.
For the past few years, banks have been focusing their initial efforts on customer experience. Personalized marketing initiatives and virtual assistants are widely deployed, but the level of advice delivered is still quite limited.
Within universal banks, the evolution of access to business data is on the critical path
Banking advice is naturally customer-centric and requires a variety of data. Within universal banks, the evolution of access to business data and the ability to link data from various sources are currently on the critical path to developing the competitiveness of banks.
The relevance of the bank's recommendations depends on: its ability to analyze the individual preferences and risk appetite of each client; the comparison of individual and collective behaviors; the consideration of a global financial situation - integrating income and banking positions, outstanding debts and the variety of asset classes held in the portfolio; the analysis of the services consumed and the linking of this information with market data
Just as Artificial Intelligence needs large volumes of data, risk modeling incorporates an ever-increasing variety of information and real-time prediction engines an increasing volume of data. To outperform their competitors, banking AI (Artificial Intelligence) needs a lot of reliable, qualified and up-to-date data.
However, the sharing of banking data has long been limited to interbank operations and legal or financial reporting. This sharing has often been a one-way process, built up over the course of regulations, business line by business line, and mostly under duress. It is, for example, the recent reinforcement of customer control requirements (KYC) that has made it possible to establish a first shared vision of the customer between the various business lines.
Driven by the search for operational efficiency and regulatory changes in the area of payments (PSD2), shared hosting projects and opening up to third parties have encouraged the implementation of information sharing infrastructures (datacenter, private cloud, datawarehouse and API, etc.).
The latest regulations on privacy (GDPR), fraud control (LCB/CFT) and banking risk (BCBS239) have necessitated the "lineage" of data on all banking activities. They have enabled banks to gain a better understanding of their data and to understand the work that needs to be done to qualify it.
The implementation of cross-functional departments specialized in Data and Cybersecurity, characterized by the appointment of a Chief Data Officer (CDO) and a Chief Information Security Officer (CISO) within each banking entity, is also a first step towards data sharing governance.
However, banking data remains fragmented, scattered within siloed application architectures, designed by business line, for a specific clientele in a specific regulatory context. Exchanges are still based on specialized systems, designed for a specific use or reporting.
However, banking data remains fragmented, scattered within siloed application architectures, designed by business line, for a specific clientele in a specific regulatory context. Exchanges are still based on specialized systems, designed for a specific use or reporting.
Two priorities emerge for a banking advice based on Artificial Intelligence to significantly support the competitiveness of a bank:
- The design of cross-functional exchange architectures and hybrid and shared data spaces designed for Artificial Intelligence (trust, timeliness, quality and volume) is on the critical path of using Artificial Intelligence as a competitive lever.
- The "Data" departments, today focused on compliance, must evolve towards the business in a recurring search for opportunity and ensure the governance of value-added data.
Banks are not so far from this. The efforts of the last few years have put them on the right track.
Almost all of them have embarked on an incremental approach, driven by the first uses but focused on data and its exploitation by Artificial Intelligence.
Most banks have been recruiting data scientists and transforming themselves for several years. For example, Société Générale already claims to have deployed 330 use cases based on Artificial Intelligence. The bank, which now claims to be "data-driven", estimates the value generated by this strategy at €230 million.
The challenge will be to stay the course in a particularly turbulent period for the markets.
The creation of a single market for data in an area where the banking industry handles large volumes of data, often highly valued
In the long term, the European Commission intends to create a market where data can flow within the EU and between sectors while respecting privacy and competition law provisions with clear and fair access rules.
Addressing the legal, economic and technical issues that cause data to be underutilized, the creation of this single data market could generate an additional €270 billion in GDP for Member States by 2028. By securing the exchange of information and access to "trusted" artificial intelligence, the EU would become a leader in a data-centric society.
The creation of this market represents a unique opportunity that banks could seize.
First, bank data is highly valued.
The emergence of open-data offers, marketplaces dedicated to data exchange, the development of "data-brokerage" and Fintechs whose revenues are mainly ensured by the exploitation of data offer new, more dynamic and more targeted observation spaces on the proven value of banking data, a secure and qualified data whose regulatory compliance is ensured.
For example, the emergence of marketplaces in the darknet makes it possible to prioritize the value of banking data.
For example, the emergence of marketplaces in the darknet makes it possible to rank the value of certain personal data. Unsurprisingly, banking data is on the podium alongside health data, but far ahead of consumer data provided by internet platforms.
Secondly, the bank has a strong experience of data purchasing and information acquisition practices - banking or non-banking - that it could value in a single data market.
Indeed, on a daily basis, banks integrate large volumes of market data from marketplaces, payment operators, specialized organizations or public services. Banking businesses are accessing increasingly diversified data sources to respond to new uses or offer new customer experiences (for example, leasing subsidiaries, wishing to move towards mobility, are acquiring vehicle usage data; private bank subsidiaries are accessing legal databases and social networks; investment banks are exploiting press data; and insurance subsidiaries are exploiting weather data from third-party government agencies).
Every bank produces and disseminates regulated information on a daily basis to financial supervisors and various institutions at the regional and international levels. The bank has a strong experience in the targeted dissemination of qualified information, including in real time.
Banks are already major players in the data market and it would be unthinkable to exclude them from it
In summary, banks are already major players in the data market and it would be unthinkable to exclude them from it, especially since, for the past few years, the strong development of online banks (e.g. in France, +30% weighted average growth[1] in the number of accounts over the period 2014-2017) and the development of associated services make them digital platforms that are sensitive to the upcoming changes in the European framework
New European regulations that could spur banks to action
Finally, while banks are not directly targeted by the draft regulations[2] governing data use in Europe, several of them could have an impact on banking activities.
The Data Act (DA) aims to allow access to personal usage data produced by connected objects or services. It appears that, to date, accounts, electronic payment instruments and smartphone applications could fall under the definition of "product" and "related services" in the text. If this is the case, the directives of the European Central Bank - and in particular the PSD2 directive - will have to be updated, and it will be in the interest of the banks to position themselves accordingly.
The regulation under study to determine the conditions of use of AI in the European Union (AI Act) provides for special management of so-called critical AI systems. AI systems used to assess creditworthiness or access to credit are concerned. In addition to the mandatory ethical and transparency obligations for all, this regulation could impose additional reporting and control measures (ex-ante) on banks, as well as special standards for training AI models and documenting processing results. The EU Liability Directive aims to facilitate the application of civil law compensation for damages caused by AI systems. This directive could also affect banks, applying to assets held by customers in the event of AI-based automatic arbitrage.
The Data Governance Act (DGA) should facilitate the sharing of data between actors in the same economic sector, between industries and with Member States. Implementing the concepts of shared data spaces described in this text may require significant standardization efforts in the financial sector. The realization of these data spaces requires large-scale work in the banks in terms of both architecture and data governance. The DGA also foresees the creation of trusted third parties in charge of collecting customer consent on data exchanges between partners. Here again, banks could create specialized structures inspired by the beneficiary declaration model they operate in the field of credit transfers.
The progressive integration of Artificial Intelligence into the advice provided by each banking business is, in the current context, an important competitive lever for banks
The progressive integration of Artificial Intelligence into the advice provided by each banking business is, in the current context, an important competitive lever for banks. It is also a powerful lever for reorganizing intra-group exchanges based on data-oriented architecture and governance.
In addition to the sectoral initiatives for shared data spaces driven by Europe, the implementation of the banking regulations relating to the single data market could encourage banks to become involved. The experience they have acquired in data exchange, mapping and flow harmonization, as well as the investments they have made in Artificial Intelligence, cybersecurity and compliance, should enable banks to take a certain leadership in the implementation of the single data market.
Banks, particularly universal banks, have an interest in this: in the short term, to satisfy the competitiveness of their activities in a profoundly modified environment; in the longer term also because banking data, by its very nature qualified and secure, should be highly valued.
François Marchessaux is a Senior Partner at Franz Partners and founding Member of Cercle K2.
---
[1] Rational: 4.4M online accounts at the end of 2017; 16.5M online accounts at the end of 2021. Sources: Bain, Franz partners.
[2] Main texts of the European data package: "Digital Market Act, Digital Services Act, Data Act, Data Governance Act, Artificial Intelligence Act and AI -Liability Directive".
---
