AI, digital banking, and new payment tools are arriving fast. Sam Boboev tracks those shifts through his Fintech Wrap Up platform and conversations with industry leaders. Sam, thanks for joining us.
Thank you for having me.
Across your coverage and interviews — what change in fintech feels most real right now?
The biggest shift I am tracking is that AI is moving closer to actual financial transactions. Before it was chatbots answering questions. Now we are getting closer to real transactions — autonomous AI agents completing banking and commerce on behalf of users. Look at what Visa and Mastercard are doing. Stripe is at every layer of payment infrastructure, as well as AI and stablecoin rails — often called the rails for AI agents to make transactions. Moving forward we will see more autonomy and agents making more transactions on behalf of us. But some questions remain open.
What do the people building these products see that the headlines miss?
The gap between a promising technology and a product that actually operates reliably at scale. I recently spoke with the chief marketing officer of Lloyds Banking Group — they serve 28 million customers, 23 million of them digital. When you are operating at that scale, even a small mistake makes a huge impact. He explained Lloyds’ human-in-the-lead approach — thinking carefully about AI because of consumer protection consequences if something goes wrong. From outside, people ask why banks are not moving fast. From inside, you see that speed is only one part of the equation.
Where are useful AI products actually emerging rather than just promises?
First, internal operations — optimising processes and efficiency within teams. Second, fraud detection — companies are actively using AI for protection, but fraudsters are also using AI to speed up their schemes. It is a race on both sides. Third — and this is the vertical I am most interested in — expense management. AI can help businesses with procurement, finding the right pricing, and executing transactions. Very promising.
What has your work on agentic commerce taught you about how AI will change payments?
The most interesting implication is that AI changes who controls the customer journey. Previously a human decided what to buy, searched for prices, chose the merchant, and approved the transaction. With agent commerce, all those steps are being passed to autonomous AI agents. An agent can find a flight, book it, make the purchase, and then book the hotel — all autonomously. This changes how we look at merchants because now they have to optimise not only for consumers but for AI agents as a new category of customer. However — JPMorgan and some of the biggest US banks have decided to restrict third-party AI agents to customer accounts because of the risks. If an agent initiates a transaction on behalf of a customer and something goes wrong, who is responsible for that loss? That question is still open. Once we solve it, agentic commerce will take off at an immense rate.
Should some financial transactions always need a person’s approval?
One hundred percent. The CEO of Papaya Global — which processes $50 billion in annual payment volumes — uses AI a lot and gives it many tasks. But when it comes to finally executing the payment, she does not trust AI. She does it herself. That tells us the most important thing: we still need humans in the loop, at least in the final execution step. Over time we will probably see more autonomy for low-value recurring transactions — subscriptions are the best place. But for high-value transactions, we still need human in the loop.
Sam, it has been great talking to you today. Thank you.
Thank you for having me.
