FSO NEWS: AI in Fintech: October 2026 News, Stocks, Rules and Use Cases

Key takeaways

In October 2026, AI in fintech shifted from chatbots to AI agents: software that can view accounts, pay invoices and place trades within limits a person sets.
Visa, Mastercard, Interactive Brokers and Danske Bank launched agent tools. JPMorgan, Bank of America, Citi and Wells Fargo moved to restrict outside agents that log in for customers.
Upcoming regulatory dates: OSFI's Guideline E-23 takes effect May 1, 2027, and EU AI Act rules for high-risk credit-scoring AI apply from December 2, 2027.
Stocks marketed on AI have had very different results. Upstart closed October 28 about 55% below its 52-week high, while Visa and Mastercard were within 4% of theirs.

AI in fintech went through a busy October 2026. Banks rolled out in-house AI at scale, card networks tested payments made by AI agents, and a large online broker began letting AI submit trades for clients. At the same time, several of the largest U.S. banks blocked outside agents, and regulators in Washington, Ottawa and Brussels set firm dates for AI oversight. This guide covers the news, the stocks and the rules, then the use cases and risks that matter whatever the news cycle.

What AI in fintech means now

For years, AI in finance meant machine-learning models working in the background, scoring card transactions for fraud or loan applications for risk. Generative AI added assistants that answer questions in plain language. The newest layer is agentic AI, meaning systems that take actions such as moving money or submitting orders, not just give advice. That shift raises new questions about who authorised a transaction and who is liable when an agent gets it wrong, and those questions drove most of this month's news.

Latest AI moves by banks, payments companies and brokers

Date (2026)	Company	AI move
Sept 17	Affirm (AFRM)	Transformer-based underwriting model live at U.S. checkout
Sept 30	Charles Schwab	Charley, a generative AI assistant, rolling out in phases from October to eligible U.S. retail clients
Oct 6	BBVA	"The Frame" model to scale AI group-wide; voice agents resolved close to 90% of interactions in pilots in Peru and Mexico
Oct 19	Santander	Three-year AI collaboration framework with NVIDIA covering models, agents, fraud/AML and real-time decisioning
Oct 19	Emirates NBD	First MENA bank to use OpenAI GPT models on in-country UAE inference infrastructure
Oct 20	Mastercard (MA)	Agent Pay for Machines opened to early access after pilots with partners including Cloudflare and Coinbase
Oct 20–21	Visa (V)	Live pilot with BILL and Billtrust in which trusted AI agents complete B2B invoice payments
Oct 21	Interactive Brokers (IBKR)	Autonomous Trading: eligible clients can let connected AI submit supported orders in a dedicated AI Partition
Oct 21	Danske Bank	Business customers can connect AI agents such as Claude or ChatGPT to their banking data
Oct 21	Bank of Ireland	AI screens about 2.9 million transactions a day for fraud; nearly 7,000 staff trained on Copilot

One theme runs through these announcements. Santander said that moving from large general-purpose models to smaller specialised ones cut serving costs in selected tests from about €5 to roughly €0.03 per interaction. Bank of Ireland said AI now screens about 2.9 million transactions a day for fraud. BBVA said 75% of its employees use AI regularly, saving an average of 2.4 hours a week. For banks, AI is increasingly a cost and efficiency story, not just a product story.

The fight over AI agent access

Not every bank is welcoming agents. According to USA TODAY, which followed earlier reporting by Axios, Citi, Wells Fargo, Bank of America and JPMorgan Chase have taken steps to limit third-party AI agents. Citi said it does not let third-party agents log in or transact, and allows only limited read-only access through approved open-banking providers. Bank of America CEO Brian Moynihan said the bank is restricting agent access to consumer deposits and wants a human approving payments. Wells Fargo's updated online access agreement, which takes effect November 24, says customers are "solely responsible" for supervising any AI agent they allow into their account.

So the industry is splitting into two models. Visa, Mastercard and Interactive Brokers are building authenticated channels with spending or trading limits, while several large banks are blocking unvetted agents until standards catch up.

Fintech stocks tied to AI

The table shows closing prices on October 28, 2026, for listed companies whose AI work was in the news. Being "tied to AI" has not protected share prices this year, and the size of a company's AI announcements has not predicted how its stock performed.

Company	AI angle	Close Oct 28	52-week range	Below 52-week high	Earnings
Upstart (UPST)	AI lending platform	$24.14	$22.21–$53.18	55%	Nov 3
Affirm (AFRM)	AI underwriting at checkout	$73.32	$42.10–$90.44	19%	Nov 5
Robinhood (HOOD)	Retail broker	$91.43	$63.52–$150.47	39%	Reported Oct 27
Interactive Brokers (IBKR)	AI trading integration	$83.54	$58.95–$98.75	15%	Reported Oct 15
Visa (V)	Agentic payments	$400.60	$293.89–$414.60	3%	Reported FQ4 in October
Mastercard (MA)	Agentic payments	$610.31	$464.52–$633.30	4%	Scheduled Oct 29

Robinhood fell after reporting third-quarter results on October 27. Revenue rose 12% to a record $1.42 billion, but diluted EPS fell 16% to $0.51 as crypto revenue weakened. Upstart, which describes itself as an AI lending platform, reports on November 3 and Affirm on November 5. Both results will test whether AI underwriting is turning into loan growth and profit.

AI regulation in finance: the dates that matter

Rule or regulator	What it covers	Key date
EU AI Act	Transparency duties (Article 50) and general application	Applied from Aug 2, 2026
EU AI Act (as amended by the AI omnibus)	New prohibitions added by the omnibus	Dec 2, 2026
EU AI Act (as amended by the AI omnibus)	High-risk Annex III systems, including credit scoring of individuals	Dec 2, 2027
OSFI Guideline E-23 (Canada)	Model risk management, including AI/ML, at federally regulated financial institutions	Effective May 1, 2027
Fed/OCC/FDIC revised model risk guidance (SR 26-2)	Replaces SR 11-7; generative and agentic AI are outside its scope	Issued Apr 17, 2026
CFPB Regulation B final rule	ECOA does not authorize disparate-impact liability	Effective July 21, 2026
SEC FY 2027 exam priorities	Emerging technologies, including "super intelligence," across 34,000+ registrants	Announced Oct 23, 2026
FINRA	Guidance on how FINRA rules apply to AI tools	Planned Q4 2026 and following

In the U.S., regulators are not writing new AI rules for banks. They are fitting AI into existing frameworks. The revised interagency model risk guidance issued in April explicitly places generative and agentic AI outside its scope, and says the agencies plan a separate request for information on AI. For lenders, the CFPB's Regulation B rule, effective July 21, 2026, says the Equal Credit Opportunity Act does not authorize disparate-impact liability. That changes how fair-lending risk is analysed for AI credit models in the U.S., but intentional discrimination remains illegal. Lenders operating in the EU will still face the AI Act's high-risk obligations.

Where AI is used in financial services

Fraud detection: models score each payment in real time using device, location, merchant and behaviour signals, and flag outliers for extra checks. Under the EU AI Act, fraud detection is specifically excluded from the high-risk credit category.
Credit underwriting: lenders use machine learning to price risk for applicants with thin credit files. Affirm said its new transformer model, live since September, increased completed purchases by 3.4% against a control group while keeping decisions explainable.
Robo-advice and portfolio tools: automated platforms rebalance portfolios and harvest tax losses, and AI now adds plain-language portfolio summaries and screening.
Compliance and KYC: AI helps check identity documents, screen against sanctions lists and sort anti-money-laundering alerts so investigators can focus on the riskiest cases.
Customer service: generative assistants such as Schwab's Charley answer account questions and handle routine tasks, with hand-off to a human when needed.
Trading: AI is used for screening, signal research and, increasingly, carrying out orders within rules the client sets.

Risks of AI in fintech

Wrong answers: generative tools can be confidently wrong. Schwab's own disclosure says Charley "may provide inaccurate or incomplete results."
Agent authority and liability: if an agent makes a payment or trade you didn't intend, your account agreement may treat it as authorised.
Bias and explainability: credit models can reproduce patterns in their training data, and lenders must still give specific reasons when they deny credit.
Fraud and security: criminals use AI too, for deepfakes, phishing and account takeover. The SEC's FY 2027 examination priorities list advanced technologies alongside fraud prevention and cybersecurity.
Model and vendor risk: banks that rely on a few AI providers inherit those providers' outages and errors.
Valuation risk for investors: AI announcements don't guarantee earnings, as the gaps in the stock table show.

Common questions

What is AI in fintech?

It is the use of machine learning, generative AI and AI agents in financial services, including payments, lending, banking, investing and compliance. Common uses include fraud scoring, credit decisions, customer assistants and trade automation.

Can an AI agent trade or pay bills from my account?

Sometimes, but only where the firm allows it. Interactive Brokers now lets eligible clients authorise connected AI to submit supported orders within a dedicated AI Partition, and Visa is piloting AI agent invoice payments with BILL and Billtrust. Several large U.S. banks, however, block outside agents from logging in to customer accounts.

When do EU AI Act rules apply to credit scoring?

From December 2, 2027. AI used to assess the creditworthiness of individuals is classed as high-risk, and the AI omnibus delayed those obligations from August 2026. Fraud-detection AI is excluded from that category.

Are fintech AI stocks a good investment?

That depends on each company's earnings, valuation and your own goals. Among the stocks in the table, distance from the 52-week high ranged from under 4% to about 55%, which shows how differently the market is pricing AI stories.

Is my data safe when a bank uses AI?

Banks remain subject to privacy, security and model-risk rules whether or not they use AI. Some, such as Emirates NBD, are using in-country AI hosting to meet data-residency expectations. Read the disclosures for any AI assistant before you share sensitive information.

Sources

This article is for general information only and is not financial advice. FSO NEWS does not recommend buying or selling any security mentioned. Stock prices are as of the October 28, 2026 close and will change. Company and regulatory details come from the sources listed and may be updated after publication. Do your own research or consult a licensed professional before investing.

This article is for general information only and is not financial, tax, legal or insurance advice. Check the official sources linked in the article and speak with a qualified professional about your own situation.