Designing the Future Financial Experience: ABN AMRO, N26 and Adyen on Product Strategy, Embedded Journeys and the Race to Own the Customer

Is Banking Becoming Invisible?

The panel opened with a provocation: is banking infrastructure fading into the background, becoming just another button inside someone else’s app? The three panelists agreed less than you might expect.

Hamerlinck argued the shift already happened, accelerated by open banking and PSD2. Banks that understood where their customers were actually spending time, rather than assuming everyone would come to a branch or a banking app, are the ones positioning themselves well. The ones still betting on branch plus website plus app as a complete strategy, he suggested, are behind.

Sisodia pushed back on the premise entirely. Banking isn’t invisible, he argued; what’s changed is the channel, not the underlying need. Customer expectations around trust haven’t softened, they’ve intensified, and they now have to be met across more touchpoints than ever. He offered a sharp example: not long ago, customers complained that chatbot responses felt too instant, too robotic. Now they’re frustrated when responses feel artificially delayed to seem “more human,” and some are asking for voice interaction instead. The specific complaint keeps changing. The demand for trust in every interaction does not.

Parikh split the difference with a distinction that shaped much of the rest of the discussion: there’s a pipeline layer and an experience layer, and they behave differently. Someone delivering food via Uber doesn’t care whether they got paid through a “banking experience”; they care that they got paid, instantly, as part of a good Uber experience. The pipeline, in his view, should be invisible — the more invisible, the better. Few institutions do both layers well simultaneously, which is precisely why infrastructure players like Adyen exist.

Who Owns Distribution When Everyone’s an Agent?

Asked which platforms will dominate distribution going forward, all three panelists landed in a similar, slightly unsettled place: agents.

Hamerlinck was candid that AI has made this harder to predict, not easier. Fewer interactions, he suggested, even within banks themselves, will involve humans; more will involve agents talking to other agents. The open question is which interface wins — embedded inside tools like Claude, Gemini, or ChatGPT, dedicated personal finance apps, or the banks and neobanks themselves. He raised, half-jokingly, the idea of “Hilton banking” — a paid tier where you can still access an actual human inside an otherwise agentic process, a kind of premium service layer on top of automation.

Parikh argued that the platforms owning the end-to-end customer experience will own distribution, pointing to the fact that no Western market has managed to replicate a WeChat-style super-app. Instead, in Europe and the US, vertical software platforms — Shopify for commerce, expense tools for corporate spend — own the relationship, with banking sitting underneath as embedded infrastructure. His view was that agents are now becoming the newest version of that same platform layer: transactions are already starting to happen through conversational AI tools, which makes those tools the next distribution channel banks need to plug into.

Sisodia, picking up a thread he returned to throughout the session, cautioned against losing sight of people in all this agent talk. Products, he argued, are ultimately built by people for people, and agents are accelerators toward an outcome, not the outcome itself. He invoked his father’s three-decade career at SWIFT — a piece of infrastructure that has been declared disrupted nearly every year for thirty years and is still standing — as a reminder that losing sight of “why” while chasing every new technological wave is how organizations end up moving fast in the wrong direction.

Partnerships: Branding Exercise or Revenue Engine?

The panel turned to partnership economics, a topic introduced with a wry nod to the steady stream of fintech-and-AI tie-up announcements that leave bank CFOs asking where the actual revenue is.

Hamerlinck noted that the word “partnership” has been stretched thin — plenty of what gets called a partnership is really procurement. He described banks balancing two pressures: the need to move fast on AI by relying on partners they’ll never build the capability to replace internally, and a growing discipline around only scaling what has a real business case, a lesson learned the hard way from over-optimistic cloud migration projects in the past decade.

Sisodia, drawing on his time at Capital One during its cloud transition in the US, agreed that this kind of transformation is expensive and slow, but argued it’s inevitable for institutions focused on tomorrow’s needs rather than yesterday’s.

Parikh offered the panel’s most concrete framework, splitting partnerships into two categories. The first is partnership for branding: high-volume, mutually beneficial for reputation, but with little stickiness and almost no margin, since either side will walk for a better deal. The second is partnership for monetization, which he argued comes from proximity to data — specifically credit data and FX data. Pricing for credit risk or FX spread, rather than just volume, is where the real money sits, and those deals tend to be the quiet ones rather than the ones announced with fanfare. In his experience, the bulk of any infrastructure provider’s partnership portfolio falls into the branding bucket, with a much smaller set of silent, data-driven deals doing the actual revenue work.

What Fintechs Taught Everyone About Product

The final theme was product design, framed around the obvious comparison: fintechs built famously intuitive, almost invisible user experiences, while traditional banks built branches and forms. What lessons actually transferred?

Hamerlinck offered a counterintuitive story. ABN AMRO once built a one-minute mortgage approval process — and customers hated it. A decision involving that much money felt wrong without some friction; the bank had to deliberately add delay back in. He extended this into a broader point about AI: agents querying which bank offers the best mortgage don’t care how a website looks. As more decisions get mediated through AI agents, he argued, banks may need to optimize less for human-facing UX and more for how discoverable and well-structured their data is to an agent doing the evaluating.

Sisodia agreed that distribution channels are shifting, but reframed the underlying goal as trust rather than frictionlessness. Not all friction is bad, in his view; what matters is whether each step in a journey — a KYC flow, a one-minute mortgage, a multi-day mortgage — builds trust or breaks it. He closed with a metaphor about fluid team roles: positions across an organization, and across a partnership ecosystem, are becoming far less rigid, and the better approach is hiring for individual strengths and motivation rather than fixed job descriptions. He recounted meeting someone who’d recently opened a long-sealed wartime bunker for the first time in decades and was demonstrating Morse code inside it — asking what motivated him revealed a passion no job title would have captured, and Sisodia argued the same curiosity should guide how teams and customers are understood.

Parikh structured his closing remarks around three pillars: trust, data, and focus. Trust, he argued, is defined less by the happy path than by how a product handles the unhappy one — how quickly and gracefully a failure gets resolved is what customers actually remember. Data is where fintechs have a structural advantage, with every micro-interaction (down to individual ISO 8583 message fields in card transactions) generating insight that can be fed back into the product. And focus means deliberately not trying to do everything: he pointed to Monzo’s early UK prepaid card, which succeeded by offering little beyond spending notifications, precisely because it knew what it wasn’t trying to be.

A Closing Word on Risk

In the Q&A, a question about regulatory exposure in bank-fintech partnerships drew a pointed response from Hamerlinck. Responsibility, he said, needs to be explicit from day one — newer fintech partners are often still maturing their understanding of where that responsibility sits, and banks typically absorb more of it. With frameworks like DORA raising the bar on third-party risk management, his advice was to stop treating it as regulatory box-ticking and start treating it as a genuine strategic capability — one of the few that, if done well, becomes a real competitive advantage rather than a compliance cost.

The session closed where it began: with the idea that the future of financial services isn’t about who builds the biggest bank, but who builds the most seamless, trustworthy customer experience — and who ends up holding the pen on that partnership.