Financial institutions are no longer asking whether AI will matter. They are trying to separate useful systems from convincing demos—and move from experiments to accountable work.
This series examines that transition from the operator’s side: what reaches production, what breaks, how governance must change, and where durable advantage will remain when capable models are available to everyone.
The essays draw on current evidence from the ECB, Bank of England, FCA, BIS, EBA, FSB, European legislation, and financial institutions deploying AI at scale.
The Model Is the Cheapest Part of Enterprise AI
Why model access is the smallest cost in enterprise AI—and why workflow redesign, permissions, controls, and adoption determine banking ROI.
The First Banking AI Winners Will Be Invisible
Banking AI creates durable wins first in fraud, operations, engineering, and advisor workflows—the valuable work that customers never need to see.
"Human in the Loop" Is Not a Governance Model
Why a person in an AI workflow is not enough: effective human oversight requires defined authority, context, triggers, and response time at scale.
When Every Bank Has the Same AI, What Still Matters?
As AI models commoditize, banks will compete on context, permissions, workflow design, trust, and the institutional ability to learn faster.
Your AI Agent Needs a Job Description
AI agents in banking need explicit goals, authority, tools, escalation rules, and accountability before they can be trusted to perform real work safely.
Your Bank Doesn’t Have a Data Problem. It Has a Permission Problem.
The hardest bank data problem is often controlled access: who and what may see which information, for what purpose, and under which conditions.
The First Wealth-Management AI Scandal Won’t Be a Hallucination
The first wealth-management AI failure is more likely to be a suitability, incentive, or accountability problem than a spectacular hallucination.
Banks Are Buying AI With Contracts Written for 2015
Conventional software contracts miss AI model drift, data rights, agent authority, exit plans, fourth parties, and provider concentration risk.
AI Will Make the Mass Affluent Worth Advising
AI can make quality advice economical for mass-affluent clients—if firms redesign the advisor workflow without weakening suitability or trust.
Banks Are Still Buying Software. AI Is Selling Them Labor.
AI is changing enterprise software from seats and features into units of work. Banks must learn to buy outcomes without surrendering control.
Working on AI inside a financial institution?
I speak with banks, wealth managers, and technology teams about adoption, partnerships, and the gap between a promising model and a working process.
Start a conversation