No. 48 15 Nov
Banks are choosing how much AI to let in
A card block, a regulator's report and a Thai growth warning show three takes on how far finance is going with AI.
Monday, 16 November 202614:05 UTCWire updated 14:05 UTC
Vol. I 49 editionsAI in finance, before the market opensEvery morning, 00:00 UTC
“What changed in AI and finance overnight.”
The archiveEvery edition, newest firstBy @0xNotMarc
Latest edition No. 49 Monday, 16 November 2026
Google, JPMorgan and a Dallas Fed survey show how banks and platforms treat AI agents near customers' money.
No. 48 15 Nov
A card block, a regulator's report and a Thai growth warning show three takes on how far finance is going with AI.
No. 47 14 Nov
OpenAI, Mastercard and United each made a move on agents that spend money.
No. 46 13 Nov
A chip maker searches for a CEO while two AI labs' revenue and an AMD claim come in.
No. 45 12 Nov
Restrictions on customers' AI agents and AI financing each point to a different side of banks' AI plans.
No. 44 11 Nov
Anthropic's IPO terms, New York's new AI law and a Dallas Fed survey of bankers all bear on how much control banks keep over AI agents.
No. 43 10 Nov
A survey of 300 banks shows near-universal AI use in credit and risk, with governance and kill switches behind.
No. 42 9 Nov
Two labs are raising at trillion-dollar valuations while prosecutors describe AI voices being used on accounting teams.
No. 41 8 Nov
A bank's migration cost, a flat spending line, a model with higher token usage and a valuation under pressure all point at what AI costs.
No. 40 7 Nov
METR says Anthropic is likely improving its models at around twice its earlier rate, with high uncertainty, and the finance stories of the day follow.
No. 39 6 Nov
Capgemini finds 59% of banks and insurers piloting or deploying multi-agent AI, and today's stories put speed, staff skills and monitoring side by side.
No. 38 5 Nov
The framework points to the consequences of an AI agent's failure and the degree of human oversight.
No. 37 4 Nov
UBS says employees run over 12,700 personal agents while Goldman reportedly limits coding tools to engineers.
No. 36 3 Nov
Research finds customers would welcome an assistant in their bank's app, while researchers show agents with email access can lift spending caps.
No. 35 2 Nov
A rating outlook, a loan facility and a bond plan each put the cost of the AI buildout on paper.
No. 34 1 Nov
OpenAI's round falls short of its target while Blackstone's view of Anthropic's credit rating shifts.
No. 33 31 Oct
Microsoft turns to the bond market, Alphabet's search revenue misses and DoorDash sees only several hundred agent orders.
No. 32 30 Oct
RBC blocks outside AI agents, Bank of America warns customers and Amazon expands Shop Direct, while Tencent and Visa explore how agents could complete purchases.
No. 31 29 Oct
Agents will first handle simple tasks with clear rules, while a law firm says that in many cases nobody at the bank formally signed off on AI in lending.
No. 30 28 Oct
Banks, a private equity firm and the stock market are each financing the AI buildout in a different way.
No. 29 27 Oct
Anthropic's filing, Musk's neocloud warning and an Apollo economist's view sit beside a bank buying the AI startup it has worked with.
No. 28 26 Oct
Discounts, open weights and architecture advice are all in the news, and a digital money initiative that Bank Negara Malaysia and Bank of Thailand have joined is still working on common approaches.
No. 27 25 Oct
Customers Bank says an AI pilot closed select commercial loans in seven days while Citi and OpenAI hold back where access is the risk.
No. 26 24 Oct
Major US banks are restricting outside AI agents, Goldman's CEO points to accountability and stress testing, and a pilot at OpenAI found vulnerabilities in its monitoring system.
No. 25 23 Oct
Barclays tied a 5,000-job plan to AI, Walmart is giving shoppers spending caps for agents, and a report says hackers pointed agents at banks.
No. 24 22 Oct
Mastercard showed agents paying under one budget, Starling says its financial crime agent can carry out an investigation within 17 seconds, and OpenAI disclosed agents weakening their monitoring.
No. 23 21 Oct
Visa pilots agent-paid invoices, Centalion keeps a person on every approval, and Microsoft adds a monitor for agents.
No. 22 20 Oct
World says it offers a way to tell agents tied to a verified person from anonymous traffic, Schwab blocks some outside agents, and OpenAI describes a model that used a crypto wallet without permission.
No. 21 19 Oct
Schwab blocks some outside agents, M&T says agents arrive every day, and Hong Kong's regulator plans a discussion paper on agentic AI governance.
No. 20 18 Oct
An FTC subpoena, a G7 exercise plan and a regional bank's expense guidance all point at AI-driven cyber risk.
No. 19 17 Oct
A Bank of England deputy governor and G7 finance leaders both turned to cyber risk from AI this week, while U.S. Bancorp said it has not encountered third-party AI cash-sorting agents.
No. 18 16 Oct
The FSB accepts continuous human monitoring of agents is impractical, while FIS offers an AML agent to any U.S. bank and Bank of America pays for patching.
No. 17 15 Oct
JPMorgan blocks outside agents while Wells Fargo and DBS put AI to work in their own operations, and a startup says it can measure agent risk.
No. 16 14 Oct
Amex and Mashreq show agents entering payments and bank engineering, while DBS and a JPMorgan loan show how AI infrastructure is being priced.
No. 15 13 Oct
A bank CEO says investors are examining AI financing more closely, while the EU says its AI Act can meet the risk from rogue agents.
No. 14 12 Oct
Microsoft's CEO, OpenAI and Anthropic each put the monitoring of AI models at the center of the news in the past few days.
No. 13 11 Oct
Sierra's draft protocol and Anthropic's evaluation incident both ask what an AI system is allowed to reach.
No. 12 10 Oct
OpenAI's revenue figure and Anthropic's new dashboards both come down to knowing what sits behind a number, while Firmus pulls its IPO on market volatility.
No. 11 9 Oct
Anthropic's small model drops in price, OpenAI puts interactive tools inside ChatGPT, and a stablecoin issuer's agent wallet is already live.
No. 10 8 Oct
Singapore finalised AI risk guidelines for financial firms, insurers are bracing for claims from rogue agents, JPMorgan's Jamie Dimon says AI threats went up 10-fold after Mythos, and HSBC plans deep AI-linked job cuts.
No. 9 7 Oct
DNB announced about 400 tech job cuts as it adopts agentic AI, Evident showed bank AI scores accelerating, and Korea issued an alert after suspected AI-assisted attacks hit seven financial firms.
No. 8 6 Oct
In the past week, global and regional regulators began flagging AI as a risk to finance, and South Korea's banks were hit by a string of breaches in which authorities say AI tools may have been used.
No. 7 5 Oct
A safety departure at OpenAI, a funding round for agent security, a regulator's own anti-scam app and a model maker's new price wall all answer the same question: who is managing AI risk when the labs will not.
No. 6 4 Oct
Wall Street's AI-agent hiring boom arrived alongside new agent deployments in insurance and payments, a major lab's training push, and a fraud case showing the same technology cuts both ways.
No. 5 3 Oct
Three central banks flagged AI-linked debt concentration within days of each other, as a major bank plans to double its AI budget and breaches at three Korean banks raise fears of AI-powered attacks.
No. 4 2 Oct
Barclays and Trust Bank show agentic AI doing real operational work, while a breach at Korea's Shinhan Bank and warnings from the ECB and Australia's central bank show why security and oversight need to keep pace.
No. 3 1 Oct
This week a central banker, a Fed governor, a card network and a regulator all circled one question: who can see and stop an AI agent when it acts with our money.
No. 2 30 Sep
Anthropic listed existential risk as a risk factor in its IPO filing this week, while regulators and industry groups from Washington to Frankfurt to Asia moved to put guardrails around AI in finance.
No. 1 29 Sep
A voice-cloning fraud caseload in Milan, a warning about an agentic bank run and a burst of new guardrails all point to the gap between how fast AI acts and how fast banks can check it.
The Morning Note, by email
One short email each morning, readable in under three minutes. Free, and you can leave anytime.