This week, the conversation around AI in banking is shifting from experimentation… to expectation.
RBC is a good example. After years of heavy investment, they’re now projecting up to $1 billion in value from AI by 2027. And in capital markets, analysts are already expanding coverage using AI — from about 1,500 companies a couple of years ago to 1,700 today, with more to come. At some point, it stops being innovation and just becomes how work gets done.
At the same time, regulators are changing their tone.
The U.S. Treasury is now openly saying that not adopting AI can be a risk — to efficiency, to resilience, to competitiveness. That’s a subtle shift, but an important one. We’ve gone from “be careful with AI” to “you probably need to be using this.”
And they’re backing that up with action. AI-driven decisions — onboarding, fraud detection, payments — are increasingly being treated like real banking decisions, which means they need to be explainable, auditable, and regulator-ready.
Meanwhile, startups are moving fast into the gaps.
Companies like Obin are automating large parts of compliance workflows with agent-based systems — delivering real efficiency gains, but also raising the usual questions around governance.
And then there’s agentic commerce — where most players say they’re ready… but integration, identity, and liability still haven’t caught up. So the takeaway this week is simple: AI is no longer optional — but neither is getting it right.
If you’re finding these breakdowns helpful, go ahead and Like and Subscribe so we can keep the conversation going.
Don’t blink — the banking singularity is accelerating.
And remember: Transformation starts with insight, so Stay Sharp and Stay Core.
RBC Thinks a Decade of AI Spending Might Finally Start Looking Like Strategy Instead of Expensive Optimism
Fresh coverage says RBC expects AI to generate CAD $700M to CAD $1B in revenue and cost savings by 2027, after already spending more than CAD $5B a year on technology and holding a top-three spot on the Evident AI Index for four straight years. Inside capital markets, AI-assisted analysts now cover about 1,700 companies, up from roughly 1,500 in 2023, with a goal of 2,500—translation: the productivity story is getting real enough that even bankers can’t keep calling it a lab experiment.
Banks are earning massive underwriting fees structuring these deals. Bridgewater Associates called this a “more dangerous phase” — the same banks writing the checks are financing companies whose AI capabilities may ultimately displace them.
Sources: American Banker
U.S. Treasury Just Opened an AI Innovation Series, Which Is Government for “Fine, We Should Probably Stop Pretending This Is Optional”
In a March 23 announcement, the Treasury Department and FSOC launched an AI Innovation Serieswith four roundtables bringing together financial institutions, tech firms, regulators and specialists to focus on practical AI deployment across fraud detection, underwriting, cybersecurity and operational risk. The bigger signal is the tone: Treasury explicitly said failure to adopt productivity-enhancing AI can itself become a risk to efficiency and resilience, while its AI Transformation Office is now pushing governance that evolves alongside deployment — basically, here’s the thing: regulators are no longer just asking whether AI is dangerous; they’re starting to ask whether not using it is its own kind of incompetence.
Sources: US Treasury
Regulators Are Starting to Treat AI Outputs Like Real Banking Decisions—Because, Awkwardly, That’s What They Are
An nalysis says nearly three-quarters of finance leaders were already using AI in 2024, 68% of institutions have increased fraud-detection spending, and 71% of fraud incidents and losses now come from unauthorized-party fraud. The important shift is strategic, not poetic: banks’ AI-driven onboarding, fraud and payment decisions are increasingly being judged under existing financial rules rather than some magical “tech exception,” which means models now need to be explainable, auditable and regulator-proof—whether that actually happens or not.
Sources: PYMNTS
Obin AI Raised $7 Million by Promising Banks an Agent Workforce That Knows What Compliance Is
Seed-funding roundup, Obin AI was reported to have raised $7M, with backing led by Motive Partners plus strategic angels including Fei-Fei Li and Lukasz Kaiser. The company says its founding team already has the trust of institutions representing more than $1T in AUM, and one deployment automated 50,000+ annual loan notices for a $500B private-credit fund, cutting processing time by 70%–90%—which is the kind of number that gets every ops executive interested right up until someone says “model governance.”
Sources: Qubit Capital
Agentic Commerce Looks “Ready” on Paper, Which Usually Means the Liability Questions Haven’t Started Yet
PYMNTS/Visa Acceptance Solutions study found 80% of acquirers say they’re ready for agentic commerce, based on a survey of 75 acquirers across Brazil, the UAE and the U.S., fielded from Jan. 16 to Jan. 30. But the same report says merchant integration costs, legacy systems, identity checks and liability rules are still slowing adoption—so yes, the rails may be ready, but “the rails are ready” and “the ecosystem is ready” are not the same sentence, no matter how confidently someone says them onstage.
Sources: PYMNTS
Don’t blink—the banking Singularity is accelerating.
If you’re finding these breakdowns valuable, give it a Like and Subscribe so we can keep the real conversations going.
