AI in Banking & Financial Services

Where AI in Banking & Financial Services actually stands

The visible movement in banking AI right now is not about models banks build themselves. It is about how they acquire capability, and the two available data points sit at opposite ends of that question. HSBC's asset management arm has taken a position in Model ML, a London-based AI startup, rather than simply licensing what it makes; terms were not disclosed. Reading that as a broader shift from vendor contracts toward equity would be premature on a single undisclosed deal. Inference

The other place things are moving is the assistant layer. Microsoft has a deal to put S&P Global's data, insights and analytics into 365 Copilot workflows and agentic experiences. The report is thin — one line, no scope, no pricing, no timing — so the useful question is contractual rather than strategic: which S&P entitlements a Copilot seat actually carries once this reaches analyst workflows. If licensed market data becomes something an analyst reaches through an assistant rather than a terminal, the value of both seats changes, and the vendor team should know which one it is paying for. Inference

It is worth being precise about what is not on this page. No supervisory or regulatory activity on AI in financial services reached this desk. No large bank published deployment figures, headcount effects, or AI-attributed cost savings. Nothing landed on fraud detection, KYC, or credit models — the areas where banks actually run AI in production. Those absences describe our coverage rather than proving the industry was still, but the loudest AI claims in this sector usually arrive via earnings calls and vendor announcements, and none of that produced anything solid enough to report.

This is the first edition of this page and it rests on a narrow base: two reported items, one of them a single-line report with an unverified label. Treat it as a starting position, not a picture of the industry. What would turn either data point into something real is straightforward to watch for — more banks taking equity where they would previously have signed a licence, or disclosed terms on the Microsoft–S&P integration showing what that data costs inside the assistant. Until then, the honest reading is that the buying relationships around AI are shifting and the deployment evidence is not yet public.