#77 - Banking’s Quiet Superpower: Making Silos Work Smarter

A pragmatic roadmap for turning banks into tech‑driven powerhouses—while keeping the risk‑proof siloes that make them safe.

Illustrated by Mary Mogoi - Website

Hi all - This is the 77th edition of Frontier Fintech. A big thanks to my regular readers and subscribers. To those who are yet to subscribe, hit the subscribe button below and share with your colleagues and friends. Support Frontier Fintech by becoming a paid subscriber🚀

For group discounts, write to me at samora@frontierfintech.io. If you can’t afford a subscription, consider referring a friend. The more you refer the longer your complimentary subscription is.

Reach out at samora@frontierfintech.io for sponsorships, partner pieces and advisory work. Spaces are becoming limited on my advisory hours. I help clients with market entry, market mapping, strategic insights, sounding board to founders and general advisory across Pan-African Fintech.

Sponsored by Paymentology

Unlock True Flexibility in Card Issuance Customers now expect richer digital experiences from their financial service providers. They want instant virtual cards, family spending controls, SME corporate expense cards, and seamless card management. The challenge? Most banks still rely on rigid, legacy card platforms that weren’t built for this level of flexibility.

Paymentology’s cloud-first infrastructure allows banks to launch agile, scalable, and customizable card programs in weeks, not months—ensuring they stay ahead of evolving customer expectations. With everything as an API, banks can now create higher degrees of personalisation for their clients.

Read about how Paycentral Leveraged Paymentology’s Infrastructure to Power its Payments Business

You will all have noticed a brand change across the newsletter. The idea was to use the Podcast launch as an opportunity to refresh the brand whilst ensuring congruence across all our media assets. Hope you guys like it.

Introduction

Jon Y

They’re all in this single well. They’re digging deep for what they’re getting at. They don’t know the other stuff well enough. In some ways, nobody knows the whole stack.

Dylan Patel

The stratification of just (Semiconductor) manufacturing is absurd. The tool people don't even know exactly what Intel and TSMC do in production, and vice versa, they don’t know exactly how the tool is optimized like this. How many different types of tools are there? Each of those has an entire tree of all the things we've built, invented, and continue to iterate upon, and then there’s the breakthrough innovation that happens every few years in it too.

Dwarkesh Patel

If that’s the case and nobody knows the whole stack, how does the industry coordinate? “In two years we want to go to the next process which has gate all-around and for that we need X tools and X technologies developed…”

Jon Y

It's a fascinating social phenomenon. You can feel it. I went to Europe earlier this year… I was talking to those people. It's like gossip. You start feeling people coalescing around something… It's a blue moon arising kind of thing. They are going towards something. They know it. Suddenly, the whole industry suddenly is like, "This is it. Let's do it."

Dylan Patel

It's like God came and proclaimed: “We will shrink density 2x every two years.” Gordon Moore made an observation. It didn’t go nowhere. It went way further than he ever expected because it’s like, “There’s a line of sight to get to here and here.” He predicted 7-8 years out, multiple orders of magnitude of increases in transistors and it came true. But by then, the entire industry was like, "This is obviously true. This is the word of God." Every engineer in the entire industry, tens of millions of people, were driven to do this.

Now not every single engineer believed it but people were like, "Yes, to hit the next shrink, we must do this, this, this and these are the optimizations we make." You have this stratification, every single layer and abstraction layers through the entire stack. It's an unholy concoction. No one knows what's going on because there's an abstraction layer between every single layer. On this layer, the people below you and above you know what's going on. Beyond that, you can try to understand, but not really...

Dylan Patel

Semiconductors are so siloed. The data and knowledge within each layer is: A) Not documented online at all because it's all siloed within companies. B) There's a lot of human element to it because a lot of the knowledge, as Jon was saying, is apprentice-master type of knowledge. Or it's "I've been doing this for 30 years," and there's an amazing amount of intuition on what to do just when you see something.

The discussion above was from an episode of the Dwarkesh Podcast by Dwarkesh Patel. He was hosting Dylan Patel of SemiAnalysis and Jon Y of Asianometry. The discussion was centred around Semiconductors and the broader energy and semiconductor economics of AI. What caught my attention in particular was this part about the stratification of the semi-conductor industry. In particular, how can it be so effective and move in unison like a hive whilst being so siloed? The questions from Dwarkesh were very clear. Across the tech ecosystem, silos are bad and are a hindrance to effective coordination. How then can things work perfectly when silos exist? Could these silos be a hidden competitive edge?

In semiconductors the silos are based on different technological elements of the stack e.g. wafer manufacturing, programming, etching and many others. In banking, the silos are based on different departments such as legal, compliance, risk, commercial and technology. In both instances, whilst stratification within the semiconductor industry is happening at a broader level, in banking it happens within an organisation. It’s by design and is arguably the reason why banks are good at risk management. For instance, credit risk being walled off from commercial teams enables them to look at credit risk independently without anchoring on their relationship with the client. Both Silos enable deep specialised skills to emerge that drive the success of the broader organisation or industry in the case of semiconductors.

Most Fintech commentators over the last 10 years have criticised the siloed nature of banks. Silos in banking have borne the brunt of every discussion about what’s holding banks back from innovating. My question was, what if silos are built into banking by design and are the secret to what makes banks successful particularly from a risk perspective?. If anything, it’s more common to hear of a Fintech shutting down than to hear of a bank shutting down. Moreover, if you speak to a number of successful Fintech entrepreneurs, an emerging consensus is that “we’re all now appreciating the value of some of the things that banks do, be it risk management, treasury or building trust in society”.

Image - Traditionally bank departments are siloed with very little communication between them.

The part that got me in particular was the section in which Jon of Asianometry says “It’s like gossip, you start feeling people coalescing around something… It’s a blue moon arising kind of thing. They are going towards something. They know it. Suddenly, the whole industry is like, “This is it, Let’s do it”. If you’ve worked with banks or around banks, then you can completely understand this. There are waves. At some point, the Oil&Gas industry is great, the next thing, avoid Oil&Gas at all costs. It’s a hive mind mentality but it could be what protects banks from wreckage.

Having said that, it’s also true that bank organisational structures are becoming an impediment to both growth and great customer service. Intelligence is knowing that two ideas that are seemingly contradictory can exist at the same time. With AI in the picture and technology continuing to play a part in how people interact with their core financial services, banks may be forced to relook at their org structures. Bank technology has evolved over many years, from mainframes to client-server core banking systems to digital banking platforms. The first two enabled banks to expand the geographical footprint but at core were enablers of the existing organisational structures. Banking was still branch-based with mobile or digital as a transactional tool. Looking forward, AI will drive the experiential layer of banking, converting unstructured data into meaningful insights and conversations for clients. I argued before that past bank innovation just shifted deterministic work onto software whilst leaving non-deterministic tasks to humans. This paradigm changes with AI. How can banks organise around these changes whilst retaining the principles of silos that make them successful?

In my recent podcast interview with Shameer Patel, he mentioned how they’re seeing high digital onboarding rates in areas where they open new branches. Moreover, these customers continue to interact with their bank digitally and never actually visit the branches. The branches are therefore no longer transactional centres and are in fact just trust markers. Therefore, does it make sense to retain the same organisational structures across branches, or would it be more impactful to organise around digital with the branch being a marker of trust?

This article will discuss how Banks and Fintechs can think through organisational structures whilst maintaining the core principles of trust and risk management that silos enabled intuitively. I argue that hybrid platform and product teams could be the solution together with actually incentivising teams to collaborate through real economics. Fintech is not just about tech, it’s about the intersection of finance and technology and part of that includes how you set up your org structures to maximise the impact of tech. Tech is but a tool.

The Context

Last week we wrote about Neobanks and specifically the top Neobanks in the world and what they get right. From Nubank’s example, we concluded that within a short space of time, it’s my bet that they’ll be the most profitable bank in Brazil and their total profits could exceed those of the incumbents that tried to block their path 10 years ago. Looking at the growth rates of other banks such as Moniepoint and Tinkoff, it’s clear that with great execution these two will soon join the special club of being as profitable as the incumbents. Interestingly, GT Bank released their annual results recently and it was clear that HabariPay whilst growing is nowhere near the performance of some of the larger Fintechs in the country. What started off as a niche sector 10 years ago is slowly becoming a major player in financial services across the emerging world. These new players are not just deploying tech, they’re also organising like tech companies with teams empowered to handle end to end customer journeys or experiences.

Source: BCG and QED

Interestingly, amongst global banks, the banks that have taken on true organisational transformation have experienced better financial outcomes over the last decade than their peers. Two banks that stand out in this regard are ING and DBS who have executed deep-rooted structural transformation projects. ING for instance, has performed strongly from a market cap perspective over the last 10 years compared to Deutsche Bank which has been a laggard. ING’s operating margin last year was 43% compared to 4.6% at Deutsche Bank and 14.91% at BNP Paribas. Similarly from a profit margin perspective, ING stood at 28% compared to 9.29% for Deutsche Bank and 16.46% at BNP Paribas. DBS as well has shown significantly better financial performance than its peers in Asia. DBS leads OCBC Bank, MayBank and UOB in terms of share price performance over the last year, year on year revenue growth and return on assets. It’s also a leader in profit margin with OCBC doing only slightly better. Whilst other factors such as market dynamics, leadership changes and strategic shifts could be the cause of these variations, over a 10 year horizon, it’s a clear leader having been named the best bank in the world by EuroMoney on many occasions. Whilst the link between market cap and structural transformation could be tenuous, it’s clear to observers that both ING and DBS have developed an agility that has enabled them to move faster than their peers.

What ING and DBS have in common is that they had a deep re-think of organisational structure with the clarity that they cannot become digital leaders by layering technology on top of their existing workflows. It’s clear that there are long-term financial benefits to shareholders from engaging in a deep rooted structural transformation.

That being said, there’s an emerging argument that banks are catching up in the app experience with Fintechs. This is a conversation I’ve had with many of my peers. Whilst not Nubank-level sleek, they’re getting good enough. On top of this, the argument is that customers don’t want much from their bank, they just want to move their money and have it stored safely. It’s a dumb-ledger the argument goes. It’s a compelling argument and is largely true. Nonetheless, if everyone becomes the same, then what economic value do you generate as an FI? When everyone becomes the same, then value will accrue to those who provide further differentiation at a rapid pace.

What will this differentiation look like? Financial services will have to move from basic transactional and store of value services to more nuanced and specialised services with AI as an enabler. A bespoke app-based private banking service to a bodaboda rider in East Africa or a shop-keeper in Nigeria. The move from deterministic digital transformation to digital services based on non-deterministic workflows is the next wave of differentiation. This will enable banks to create additional and defensible value. We’ve seen across the world that more and more of the value in AI is accruing to the app-layer. This will be the case in financial services.

To achieve this new vision, banks and FIs in general will have to rethink their organisational structures from a first principles perspective. It simply won’t cut it to layer tech on top of the existing structures. There’s an economic imperative as can be seen by the long-term performance of ING and DBS. There’s also a technological imperative to drive further differentiation between FIs and their competitors. Jamie Dimon supports this view as he said in his 2021 annual letter “ Banks ... are facing extensive competition from Silicon Valley, both in the form of fintechs and Big Tech companies (Amazon, Apple, Facebook, Google and now Walmart), that is here to stay. As the importance of cloud, AI and digital platforms grows, this competition will become even more formidable.” He added in the same letter that “It is completely clear that, increasingly, many banking products, such as payments and certain forms of deposits, among others, are moving out of the banking system. In addition, lending in many forms – including mortgage, student, leveraged, consumer and non-credit card consumer – is moving out of the banking system.” It’s a threat that can’t be ignored.

The Challenge

I will go through some of the contextual challenges that banks face when it comes to innovating on top of existing organisational structures that at their core are based on product marketing. This is the idea of creating products and driving their sales through branch based structures.

1. Whilst a number of banks are changing, most digital products are sold through the existing branch based product marketing structure. In this case, innovation happens centrally when it comes to product design and ideation whilst existing distribution sells this innovation. There are a number of issues with this;

   1. From a tech perspective, whilst the software is being built using agile methodologies, the customer feedback and iteration is happening in a traditional manner. You then get a situation where people at the branch level get feedback that is not intuitively flowing back into the centralised software teams;