Discussing Banking and Wealth Themes for week ending Sept 9, 2023

Banking: The End of Globalization

Given a continued absence of transactional data to evaluate this week, I thought I would continue to explore evolving themes across banking and wealth management. In the former, one that I continue to witness is global retrenchment. I remember back in the 90’s and early 2000’s (i.e. pre-financial crisis) that perhaps 20-30 banking groups were pursuing, with various speed, and operational intent, global infrastructure development.? This was particularly focused on, either through acquisition or application building a banking footprint in the most promising emerging global economies of scale. While most saw this as the way that a banking group could best align itself to the needs of global multi-nationals, others saw themselves as truly creating domestic brands that would serve retail clients as well as the most promising SMEs developing themselves to challenge less advanced operational and technically advanced domestic banks.

Fast forward, post the financial crisis, and the regulatory changes it has introduced around capital provision, and ringfencing local banking licenses in order to reduce “contagion” risk, and one sees that globalization trends are being aggressively unwound, being usurped by either integrated vertical strategies, or by more manageable regional ambitions (confined to South and Central America). This has been particularly noticeable in the disposal of late by Citibank, and HSBC, but it has even been included firms with very specific global agendas such as Standard Chartered, through some of their recent announced disposals in Africa.

Assuming they can get regulatory approval (which is not always guaranteed), one would expect the clear winners of this will tend to be large (but not the largest) domestic players who can (possibly) acquire economies of scale with substantial synergies to keep their cost/income ratios in check, even perhaps creating opportunities for new types of segmentation-oriented banking.? However, there is also an argument to be made that ambitious fintech participants in the payment and lending arena, should their valuations improve again back to 2021 heights, who seem intent on expanding their footprint to include certain types of banking services, could also emerge as acquirers.

In any case, there seems be no clear path back toward a resurgence in global ambitions, and thus I am left briefly to ponder exactly what multi-nationals, who are still being served via the corresponding bank model, and by many banks retaining more limited corporate banking capabilities, will do, and whether some of the “creeping” we are witnessing is a sign of more to come.? When one thinks about where the largest multi-nationals are increasing “intruding” into certain functional segments of the banking services market, one can’t help wondering when one or more of these players, either individually or through a new “entity” or network, decides to make a bold play at disrupting the corresponding banking model and replace it with something that is more capital and infrastructure light in nature, but yet effective in servicing “cross-border” and domestic supply chain needs.? I don’t think any of the dozen or so firms that we all “can name” has yet articulated a blueprint with this in mind, but the continued retrenchment, perhaps combined with truly resilient blockchain based stablecoin platforms, will be the catalyst.

Wealth Management and AI:? Tie Ups & Productivity

Continuing with the topic I discussed last week, I noted, during the last seven days, with more regularity, announced tie-ups between companies with mature data science and artificial intelligence capabilities and financial firms in the wealth management market.? While the announcement themselves focused a lot on the accessibility to data science and advanced model delivery, the tie-ups also underscored the fact that if AI firms want to successfully translate capabilities into commercially valuable and repeatable use cases that can be supported by precise business process, and logic, they need large integrated wealth firms, or networks, as much as these organizations need them.?

This should have logically led to a lot more announced partnerships, but this hasn’t happened quite as quickly as I, nor I imagine, anxious venture capitalist, expected for a number of reasons. First, the AI landscape of potential partners remains noisy and by extension, confusing.?? There are certain market leaders that have particularly strong capabilities in specific AI modelling techniques, esp. in the semantic and computer vision areas that often represent safe options for large enterprises, but these firms can often require long lead times to turn general purpose capabilities to fit domain rich areas of interest.? Thus, many financial firms often feel compelled to interact with both general-purpose firms as well as more sector / domain focused organizations. This creates a much lengthier period of investigation before any actual useful experimentation can begin.

Second, many financial firms have wrongly operated under the premise that with the right data science expertise, combined with a tightly defined value proposition, as well as a mature data management capability, one can quickly leverage a diverse range of machine learning techniques to acquire compelling results. However, this has so far proven untrue, as the industry has come to realize that data wrangling and precise data labelling is actually quite difficult, not to mention, trying to apply this to the appropriate model that delivers meaningful results for particular use cases.?

Third, and finally, financial firms have come to realize that applying AI, especially when deep learning and transference methods are applied will only be adopted and trusted if it is explainable as well as recognizable. This doesn’t mean it needs to adhere to the confines of human drawn insight but applying only the highest weighted probabilities to multi-dimensional outputs, doesn’t automatically satisfy the “sniff” test.

The truth of the last two points should be reassuring to fintech oriented AI organizations assuming that they have reached a level of maturity in the past five or six years where they have not only thought through the value generation from fulfilling one or more key business cases that transform productivity, efficiency and client outcomes (for example), but also on how to properly develop the data management infrastructure to make the AI models they seek to leverage, consistently explainable.? I have often seen that AI firms are always trying to rapidly prove they can rapidly leverage their model capabilities as well as data classification skills (when faced with unstructured but definable datasets), but partnerships only seem to evolve and flourish when the sometimes-elusive notion of explainability becomes evident.

I am not sure if the fact that the seasonal lull in the summer months with deal and partnership activity is starting to move behind us that has prompted more deals to be announced than normal where enterprise and networks are seeking out AI partnerships, but I don’t think anyone should be surprised if even very large data scientist enabled organizations start to broaden and deepen their AI partnership ecosystem to both gain competitive advantage as well as achieve consistent, explainable results.

?