エンベデッド・ファイナンス、デジタルID、およびChatGPT：MONEY 20/20 EUROPEからの考察

Money20/20 – both the European and USA versions – are among my must-attend events, as they never disappoint. They remain the best way to catch the zeitgeist in banking, fintech, and payments across both sides of the Atlantic. This year, Kieran Hines and I represented Celent in Amsterdam, and there was a strong contingent of our colleagues from Oliver Wyman.

Again, we had a couple of very busy and productive days meeting with clients and prospects, while trying to catch a few content sessions in between. Inevitably, any reflections from such a big and broad event are going to be shaped by personal interests. Kieran already shared his observations on where open banking and embedded finance are headed, and I wanted to add a few thoughts as well.

Embedded Finance and BaaS: Maturing Approaches

In the US, a company with BaaS technology typically needs a banking partner who brings the necessary licenses, such as BIN sponsorship. Of course, in Europe, that’s not always the case – technology firms can obtain and operate under different licenses, such as EMI (Electronic Money Institution), enabling them to offer some services without a bank. Despite these differences, there have been instances where BaaS relationships failed or came under scrutiny in both the US and Europe, prompting the players to focus on stronger governance and partnership management. The regulators are also taking note; in Europe, the upcoming PSD3 is expected to bring “further legislative inclusion” of EMIs.

Banks, of course, still have a role to play in embedded finance. Last year, NatWest announced a partnership with Vodeno to launch NatWest Boxed, a BaaS business. The fireside chat between Andy Ellis, CEO at NatWest Boxed and Kim Van Esbroeck, Chief Revenue Officer at Vodeno was standing room only, and deservedly so. It articulated very clearly both the opportunity for BaaS – not just large e-commerce/ tech brands and SMEs, but also banks in other markets – and what it takes to capture it, with designing and delivering a seamless customer experience across partners high on the list of critical success factors.

Digital Trust: Towards the New Digital Identity Wallet

I had several very interesting conversations around the broad topic of digital trust. This remains a rather fragmented space with many specialist solutions tackling different parts of the problem, such as using digital footprint to prevent application fraud, helping clients screen for sanctions, or enabling electronic signatures compliant with local regulations.

Many also talked about the upcoming EU Digital Identity (EUDI) Wallet. It’s an ambitious undertaking and there are four large scale pilots underway to test how the wallet might support various different scenarios. With many Money20/20 interactions being more of a “meet-and-greet”-type, I now have a few briefings set up to explore this topic in more detail. Look out for a research report from us on this over the summer.

ChatGPT: A Quiet Before the Storm?

While AI has been a big theme at the event, ChatGPT and other generative-AI technologies weren’t as prominent as I would have expected. Of course, as somebody pointed out to me, for such big events, much of the conference agenda gets firmed up months in advance, and these technologies haven’t really come to the fore until December or so last year.

Still, I managed to go to a great session called “Harnessing the Hype” with ABN Amro and Citi executives discussing what generative AI means for them. Not surprisingly, both institutions are taking it seriously, calling it a “board-level topic” and “the most important thing of our careers”.

I was particularly intrigued by a couple of early use cases shared by ABN Amro, both in the contact centre. First, the large language model (LLM) tool is listening to the agent’s conversation with the client and capturing the notes into the CRM system. The agent is freed up to focus on the conversation and only needs to review the notes. Second, like in many large banks, the contact centre agents are surrounded by multiple screens and often have to navigate multiple pages to access relevant client details. Again, the LLM hears the client question and brings up the relevant information automatically, saving the agent time.

Of course, both institutions spoke of the importance for strong governance and control mechanisms for trying out new technologies. For example, LLMs are typically deployed in contained environments, with no bank or client data used for training the overall model. I expect we’ll hear a lot more about generative-AI in Las Vegas!

Finally, I’d like to thank everyone that met with us. We know that time at Money20/20 is a precious commodity, and we appreciate your generosity. For us, every conversation is valuable, as we get to know better what’s on your mind and how we can best serve you. We look forward to continuing the dialogue!