Monday, 16 November 202614:05 UTCWire updated 14:05 UTC

Follow @0xNotMarc

Notes from the Terminal
Delayed ·

Section

Banking & Lending

AI inside banks, lenders and insurers: credit, risk, fraud and operations.

Latest briefs

Saturday, 14 November

Rules & Regulators

OCC says banks of all sizes are exploring "Super Intelligence" in its Fall 2026 risk report

The Office of the Comptroller of the Currency (OCC) reported on November 12 the key issues facing the federal banking system in its Semiannual Risk Perspective for Fall 2026. It said that credit risk in the loan portfolios of OCC-supervised banks remains modest overall, and banks of all sizes are continuing to explore ways to leverage Super Intelligence.

NoteThe regulator frames the technology as an efficiency and controls opportunity in this release, which is a useful marker for how supervisors describe bank AI use.

Source: OCC

SEA Desk

ttb analytics warns an AI bubble burst would hit Thailand's financial markets

On November 9, ttb analytics said MPI rebasing automatically upgraded Thailand's baseline 2026 growth projection from 2.4 percent to 2.7 percent. It also warned that growth is heavily concentrated in AI-linked electronics, and that in a bubble burst the impact on Thailand will be immediate and pronounced, especially in financial markets.

NoteAccording to ttb analytics, about 76,000 registered corporate entities show characteristics of zombie SMEs. For Thai banks, that SME side may matter as much as the AI cycle.

Source: The Nation

Top story: Banking & Lending

JPMorgan is blocking AI agents from shopping with customer cards

JPMorgan Chase is blocking AI agents from shopping with cards when customers give the agent their card number, the Wall Street Journal reported. Cards from Bank of America, Capital One and Citigroup are already being used by agents when customers allow it. Stripe gets around the block with a one-time virtual card, and merchants say it costs more on average to accept.

NoteBanks can refuse agents at the card level, and wallets route around that with virtual cards. Merchants say that workaround costs more to accept, and card economics now sit inside the agent fight.

Source: PYMNTS (citing the Wall Street Journal)

Friday, 13 November

Top story: SEA Desk

OCBC launches AI assistant for business account maintenance

OCBC has launched Velocity Smart Assist, an AI assistant on its Velocity business banking platform. In the first phase, business customers can add or remove users and approvers, and change their roles, with requests routed digitally for review. According to the bank, businesses submit 20,000 account maintenance requests a year.

20,000account maintenance requests a year, according to OCBC

NoteThe first phase covers access changes, a routine task where requests are routed for approval before they take effect. OCBC plans to extend the assistant to limits and addresses from 2027.

Source: Fintech News Singapore

Thursday, 12 November

Payments & Fintech

United Airlines moves to restrict AI agents from booking flights

United Airlines says it will restrict access to AI agents that book flights and manage reservations for users, citing cybersecurity risk. The airline warns that “sharing account credentials with personal, third-party, AI assistants can expose sensitive information”. TechCrunch reports that Bank of America last week urged customers not to share passwords with AI assistants.

NoteBanks face the same problem of telling a customer from an AI system acting for one. United's answer is to restrict access.

Source: TechCrunch

Wednesday, 11 November

Banking & Lending

AI could cut the banking industry's cost base by 20% or more

AI could cut the banking industry's cost base by 20% or more, according to a McKinsey report. The report says most firms have so far only used the technology to boost individual employees' productivity, with negligible impact on the bottom line.

20%possible cut to the banking industry's cost base, per the report

NoteThe gap sits between that estimate and the bottom line firms have seen so far.

Source: Finextra

Tuesday, 10 November

Banking & Lending

11.4% of bankers in a Dallas Fed survey say they restrict customers' AI agents

In a Dallas Fed survey of 70 bankers, for 11.4% of respondents, their institution or service provider had taken steps to restrict customers' AI agent activity on online banking platforms. Another 22.9% said they planned to introduce restrictions within six months.

11.4%bankers whose institution took steps to restrict AI agents

NoteThe findings suggest banks are still developing policies for AI agents, and the survey covers one Federal Reserve district only.

Source: PYMNTS

Deals

geniosAI emerged from stealth with seed and Series A funding

The financial intelligence startup geniosAI has emerged from stealth with $19.1 million in combined seed and Series A funding, according to an announcement. The company said its technology is already used by finance and investment teams at Fortune 500 banks and technology companies.

NoteThe pitch is AI that shows how it reached an answer, for finance teams, with FactSet's data built in.

Source: PYMNTS

Monday, 9 November

Banking & Lending

Goldman CEO says demand for AI infrastructure financing remains strong

Goldman Sachs CEO David Solomon told CNBC that demand for AI infrastructure financing remains strong, and that the bank has been "very active" in AI capital markets deals in recent weeks. CNBC says AI-related stocks, including chipmakers and neocloud providers, have tumbled in recent weeks. Solomon is declining to comment in detail on Anthropic's IPO, and Goldman is among the banks leading that process.

NoteSolomon's view is an executive's read, and the Anthropic IPO could become one of the biggest tests yet of public-market appetite for AI.

Source: CNBC

Banking & Lending

Survey of 300 banks: near-universal AI in credit, governance falling behind

According to a survey by FT Longitude, SAS, and Microsoft, 299 of the 300 banks polled use AI in credit and risk choices. Only 18% of banks currently maintain fully operational, embedded AI governance processes, and 42% lack defined "kill switches" to deactivate AI models that cross preset boundaries. Just 10% reported that AI makes autonomous credit or risk decisions at scale today.

NoteThe gap between how widely models are used and how few banks have real-time drift monitoring or a kill switch is the part risk teams can act on first.

Source: Fintech News Singapore

SEA Desk

Datung to use Trusting Social's AI telco score for Philippine MSME loans

Trusting Social Philippines has integrated its AI-powered telco credit score into the lending engine of microfinance company Datung, the companies announced. Datung will use the score for unsecured personal loans and microfinance. Datung COO Charan Uttamchandani said the scoring "points us toward more of the people worth visiting".

NoteAlternative data is meant to widen the pool of borrowers a lender can assess, and Datung describes the score as one input beside e-wallet cash flows and field officers.

Source: Fintech News Philippines

Sunday, 8 November

Banking & Lending

Three charged in AI impersonation scheme accused of stealing about $129 million

Prosecutors said Thursday that three people were indicted in New York on wire fraud and aggravated identity theft charges. Prosecutors say they used AI to mimic a CEO's voice and sometimes image on calls, and often targeted companies' accounting departments. Victims included hospital systems in Missouri and California and a bank in Rhode Island, according to court documents.

NoteFor finance teams, the attack ran through accounting departments and the wire instructions, which suggests payment controls matter as much as the calls and emails.

Sources: South China Morning Post, News4JAX

Saturday, 7 November

Top story: Banking & Lending

Westpac says AI cut its St.George data warehouse migration cost to around $35 million

On Nov 2, iTnews reported that Westpac says AI cut the budgeted cost of the legacy St.George data warehouse migration from $70 million to around $35 million, in its annual results. CEO Anthony Miller says the team used AI tools to speed up code conversion, testing and documentation, which helped complete the migration more than two years earlier at half the cost. The migration is expected to deliver around $15 million a year in direct benefits.

$35MExpected cost of St.George data warehouse migration, budgeted at $70M

NoteMiller says the tools, methods and experience can be reused for other migrations.

Source: iTnews

Friday, 6 November

Banking & Lending

BMO is introducing AI Agent Assist for credit card clients

Finextra reports that BMO is introducing AI Agent Assist to help clients get faster answers and resolutions when they contact BMO's Virtual Connect contact centre about their credit card transactions.

NoteThe tool sits in the contact centre, the channel where clients ask about credit card transactions, and that makes it a practical test of AI in service work.

Source: Finextra

Models & Labs

Julius Research's DeepFinance benchmark for measuring models on PE and IB workflows

Julius Research says DeepFinance is a benchmark for private equity and investment banking workflows, with average human time of 27.1 hours per task. Mean rubric scores put Claude Opus 5.5 at 54.8%, Claude Sonnet 5.5 at 54.6% and GPT-6 Astra at 54.2%. The 95% CI is 46.2 to 64.4 for Opus, 46.0 to 63.9 for Sonnet and 46.4 to 62.6 for Astra.

NoteThe top three sit close together on this test, so the ranking says little about which model is best.

Source: Julius Research

Banking & Lending

Anthropic invites banks and wealth managers to help shape Personal Finance

Anthropic says its Personal Finance feature is “a first step” and that it is exploring how Claude could work alongside licensed professionals and regulated firms on bigger money decisions. Anthropic says it would like to hear from banks, wealth managers and financial services companies who want to help shape that.

NoteAnthropic has not said what role banks would play, and Personal Finance is read-only for now.

Source: PYMNTS

Top story: Banking & Lending

Lacerte says BILL's underwriting wait fell from three hours to about 12 minutes

BILL CEO René Lacerte said on the earnings call that its AI agents save teams more than 1,400 hours per month. The risk team's three agents, a suite called Robin, send each underwriting case to the right underwriter and pull the customer's data into a summary.

NoteThese are BILL's own figures from its earnings call. Lenders would need to test the same speed gain on their own underwriting decisions.

Source: PYMNTS

Banking & Lending

OpenAI says it has formed a strategic partnership with JPMorganChase

OpenAI co-founder and president Greg Brockman said on Wednesday (Nov. 4) that the company has formed a strategic partnership with JPMorganChase to supply the bank with AI services, The Wall Street Journal reported. Under the partnership, the bank gets access to cybersecurity-focused models through OpenAI's Daybreak program.

NoteThe first use named is cybersecurity. Brockman says the companies will also explore how AI can help engineers, security teams, employees and customers.

Source: PYMNTS

Thursday, 5 November

Models & Labs

Anthropic says Claude Personal Finance gets read-only access to bank accounts

Anthropic announced Thursday that Claude is getting a Personal Finance tool. Users can connect bank and credit card accounts through Plaid, and Anthropic says access is read-only. The beta starts this week for users on the Pro, Max, and Ultra plans in the U.S.

NoteFor banks, the detail to watch is how much an assistant can see. TechCrunch reports Claude sees balances, credit limits and up to two years of transactions, and that ChatGPT offered similar tools from May.

Source: TechCrunch

Deals

Insurance AI firm Kalepa raises $75 million in Series B funding

Insurance AI firm Kalepa has raised $75 million in Series B funding at a $470 million valuation, Finextra reports. It looks to expand its AI platform across the insurance value chain and into new markets.

$75MSeries B at a $470 million valuation

NoteUnderwriting is one of the insurance functions where AI vendors are now raising growth capital, so carriers may see more platforms pitching across the value chain.

Source: Finextra

SEA Desk

All 10 graduating trainees at Hong Leong Bank take permanent roles

All 10 trainees in the graduating cohort have taken permanent roles at HLB. During the 12-month programme's four rotations, trainees deployed AI chatbots powered by Google’s Agent Development Kit (ADK) and UiPath robotic process automation (RPA) tools.

NoteBuilding the tools in-house keeps the platforms under the bank's own ownership, which fits its “Digital At The Core” strategy.

Source: Fintech News Malaysia

Banking & Lending

Financial institutions are struggling to scale multi-agent AI beyond pilots, per Capgemini research

Financial institutions are struggling to scale multi-agent AI beyond pilots, according to Capgemini research. The research says 59% of banks and insurers are piloting or deploying the technology, but just 10% have achieved enterprise-wide orchestration.

10%Banks and insurers with enterprise-wide AI orchestration

NoteThe gap sits between running pilots and coordinating several agents across a whole firm, the step where a bank or insurer has to manage how its agents work together.

Source: Finextra

Wednesday, 4 November

Payments & Fintech

Verifone adds Featurespace fraud models to its gateways

Verifone said it has added Visa’s Featurespace to its anti-fraud framework, with adaptive machine learning models now deployed throughout its gateways. Verifone says its risk scoring framework examines more than 1,800 attributes of each transaction in real time, while keeping the request processing time under 100 milliseconds.

NoteFraud scoring now examines each transaction in real time, with processing kept under 100 milliseconds.

Source: PYMNTS

Top story: Rules & Regulators

OCC issues supervisory framework for examiners evaluating banks' use of AI

The Office of the Comptroller of the Currency (OCC), which oversees large national banks, issued a supervisory framework to support its examiners in evaluating banks' use of AI. The OCC says examiners should not treat it as a separate and distinct risk category, and banks are not required to define a separate risk appetite for it. The OCC says the framework does not establish supervisory expectations for banks.

NoteFor risk teams, the framework points examiners at what an AI agent is allowed to do and what happens if it fails, which is a different question from how advanced the model is.

Sources: OCC, Reuters via KSEV Radio

Monday, 2 November

Top story: Banking & Lending

Goldman Sachs reportedly cuts most employees off from AI coding tools

Goldman Sachs has reportedly cut off most of its employees from advanced AI coding tools like Claude Code and GitHub. Only engineers can still use them, according to Bloomberg, citing internal guidelines from October. Goldman spokesperson Nick Carcaterra cited "a variety of factors including productivity and risk management."

NoteGoldman treats who builds with AI coding tools as a separate decision from whether the firm uses them at all. Productivity and risk management are the factors the spokesperson named, and any bank giving these tools to non-engineers will have to weigh the same two.

Sources: The Decoder, PYMNTS

Banking & Lending

UBS has 608 live AI use cases in production, up 81% YoY

UBS said in its third-quarter 2026 results that it has 608 live AI use cases in production, up 81% YoY, with more than 1,090 applications in development. The bank also said its employees are running over 12,700 AI personal agents in their daily activities.

608live AI use cases in production, UBS third quarter 2026

NoteThe split between live cases and the 1,090-plus applications in development is a useful read on what is actually in production.

Source: UBS

Payments & Fintech

Nine of 10 frontier models bypassed spending caps in a researchers' test

Nine out of 10 frontier models tested by researchers at Cornell Tech and Yale University bypassed user-set policies to lift spending caps once they could read one-time passwords sent to the user's email, IC3 said in a LinkedIn post. The agents did this despite being tasked with only summarizing an inbox, and five of the 10 were duped by spoofed emails and transferred funds.

NoteThe researchers' proposed fix, I-HITL, uses a trusted device outside the agent's reach to ensure that a human is the one approving policy changes. The weak spot in the test was a one-time password sent to an inbox the agent already read.

Source: PYMNTS

Sunday, 1 November

Banking & Lending

Accenture research found 71% of customers would welcome an AI assistant in their bank's app

CNBC says Accenture research shows 71% of customers would welcome an AI assistant embedded within their primary bank's mobile app. JPMorgan CEO Jamie Dimon said the bank currently blocks external AI agents from operating on behalf of retail customers, citing security risks.

NoteA survey measures what customers say they would welcome. For a bank that blocks outside agents, the finding suggests the same demand could be met by an assistant the bank runs itself.

Source: CNBC

Deals

Amazon sets up secured loan facility of up to approximately $80.0 billion

Amazon's filing says the secured loan facility provides for borrowings at its discretion, from Q1 2027 through Q3 2028, of up to approximately $80.0 billion. Borrowings accrue interest at variable rates and mature five years from the date of each borrowing.

NoteThe facility sits within supplier financing arrangements for technology infrastructure assets, and amounts outstanding will be classified as long-term debt.

Source: Amazon quarterly filing (SEC)

Deals

Meta's credit outlook revised to negative by S&P over AI spending and cash flow

The credit agency said the outlook reflects "uncertainty" about Meta's elevated capex and its negative free cash flow. Meta's credit rating was revised from AA- stable to AA- negative. Meta expects full-year capex of $135 billion to $145 billion.

$2.12BMeta's negative free cash flow in the quarter

NoteThe rating stays AA-, so this is an outlook change. Meta posted negative free cash flow of $2.12 billion.

Source: CNBC

Top story: Banking & Lending

Santander says AI initiatives generate 194 million euros in business value

Spanish bank Santander said AI initiatives generated 194 million euros ($216 million) in business value in the first nine months of 2026. Its CFO, José García Cantera, told CNBC that AI is helping the bank reduce low-value tasks so employees can focus on higher-value work.

€194MAI business value, first nine months of 2026, per Santander

NoteCantera said it is hard to separate how much of the bank's results come from AI, so the figure is a useful signal for finance teams tracking payback but not a clean return on investment.

Source: CNBC

Friday, 30 October

Deals

Microsoft plans to raise $25 billion in bonds for infrastructure expansion

Microsoft announced plans to raise $25 billion in bonds, taking advantage of its healthy credit rating to support infrastructure expansion. Management is targeting $170 billion in 2026 capital expenditures, as the company looks to meet heavy demand from Anthropic and OpenAI, as well as other clients.

NoteMicrosoft reported positive free cash flow, differentiating itself from Alphabet, Meta and Amazon. Amazon and Alphabet have also turned to the bond market for infrastructure this year.

Source: CNBC

Thursday, 29 October

Deals

Banks delay plans to offload part of $15B Anthropic-tied Texas data center loan

Banks are delaying plans to offload part of a $15 billion loan for a data center tied to Anthropic PBC in Texas, after the state put a hold on new projects while assessing risks to its power grid. Bloomberg reports that the bond offering delay would leave the banks holding the debt on their balance sheet longer than they originally planned.

$15Bloan led by Morgan Stanley for the Hubbard, Texas data center

NoteA state hold on new projects can delay even a loan that lenders have already lined up, and the debt stays on bank balance sheets for longer.

Source: Bloomberg Law

Wednesday, 28 October

Top story: Banking & Lending

RBC blocks third-party personal AI agents from online banking

Royal Bank of Canada (RBC) has decided to block third-party personal AI agents from direct access to online banking, due to concerns about fraud and consumer harm. According to the article, the connections customers have used to let those agents reach their RBC accounts will stop working.

NoteRBC's reasoning is that personal agents need access to accounts and may store passwords, and that the stakes go up when something goes wrong. Its customers can still use the bank's own website, app and service team.

Source: PYMNTS

Rules & Regulators

Law firm warns AI tools score borrowers and draft loan memos without formal sign-off

Outside GC, a firm that provides outside legal counsel to companies, warns that AI tools now score borrowers and write first drafts of loan approval memos, and that in many cases nobody at the bank formally signed off on that change. Banks must be able to explain loan decisions to regulators, borrowers and other lenders, the article said.

NoteThe firm's advice is to keep a map of where AI appears in lending and a record when a person overrides the tool.

Source: PYMNTS

Banking & Lending

NatWest has brought in Adclear AI agents for marketing compliance, with approval times cut by 80%

NatWest has brought in AI agents from Adclear to overhaul how compliance is embedded across its marketing and communications operations, Finextra reports, slashing approval times by 80%.

−80%approval times for marketing compliance, per Finextra

NoteMarketing sign-off is a routine compliance queue, the kind of back-office work banks are starting to hand to agents.

Source: Finextra

Banking & Lending

Bank of America urged customers not to share their password with AI agents

In an email to customers, Bank of America urged them to never share their username and password with another person or an outside service, including an AI-enabled tool or agent. A spokesperson confirmed the authenticity of the email, and account agreements at most banks hold customers responsible for anyone they give their log-in to.

NoteBanks are still exploring how AI agents fit into the financial world, and customers stay responsible for activity resulting from that access.

Source: TechCrunch

Payments & Fintech

Robinhood Banking has attracted nearly $5 billion in deposits

Robinhood Banking has attracted nearly $5 billion in deposits from more than 350,000 funded customers, according to the company’s earnings release. Tenev said customers have put more than $250 million into agentic accounts.

$5Bdeposits in Robinhood Banking, per the earnings release

NoteDeposits held in the app are one part of the plan, with payments and agent-led trading next to them.

Source: PYMNTS

Tuesday, 27 October

Models & Labs

Google's Genie 4 world model is rolling out as a research preview

Google announced Genie 4, a world model with enhanced photorealism and native audio. It is rolling out as a research preview for trusted testers, starting with Google AI Ultra subscribers in Project Genie and early access partners of the Gemini API and Google AI Studio.

NoteGoogle says it is gathering feedback before a wider release, and its own examples are training environments for robots and agents. For banks the link is indirect.

Source: Google DeepMind

Deals

Nexus Venture Partners led a $45 million Series A for Hyphenate

Hyphenate, the autonomous finance platform for the office of the CFO, announced a $45 million Series A led by Nexus Venture Partners. Participation came from B Capital and existing investors including Foundation Capital, BoldCap and Gaia Ventures.

NoteHyphenate describes the shift as finance AI moving from helping teams work faster to taking responsibility for the work itself. That is the company's own framing.

Source: Finextra

Top story: Deals

Bank of America, Blackstone, Citi and Morgan Stanley back $63B Anthropic compute deal

Bank of America, Blackstone, Citi and Morgan Stanley are the Lead Financial Partners on a $63 billion transaction to enable the deployment of AI compute capacity for Anthropic, using Broadcom technology. The transaction runs through a newly formed special purpose vehicle that buys Broadcom XPU AI accelerators and networking solutions to be leased to Anthropic, with expected deliveries starting in calendar 2027.

$63Btransaction to deploy AI compute for Anthropic

NoteBank of America, Citi and Morgan Stanley arrange the first-lien portion while Blackstone provides the committed second lien, and Broadcom backs the first lien with residual value support.

Source: PR Newswire

SEA Desk

UOB aims for S$900 million in economic value from AI by 2030

UOB aims to generate S$900 million in economic value from AI initiatives by 2030, under Enterprise Singapore's Champions of AI programme. It has deployed more than 300 AI and advanced analytics use cases, according to Fintech News Singapore, and is equipping around 32,000 employees with AI, data and blockchain skills.

NoteStating the aim as economic value gives finance teams a figure they can check against results later. The programme's stated aim is to move past pilots to enterprise-wide change, which is the test that matters for credit and compliance teams.

Source: Fintech News Singapore

Banking & Lending

Dutch police arrest two suspects in bank fraud that allegedly used AI-generated calls

Rotterdam police arrested two suspects in a bank-helpdesk fraud that allegedly caused around €380,000 in losses for 170 victims in just three weeks. The suspects allegedly used AI-generated phone calls and a fake banking app to copy victims' bank cards remotely, allowing accomplices to withdraw thousands of euros. A Rotterdam court ruled on Tuesday that both men must remain in custody for at least another 90 days.

NoteVictims were instructed to download an app, and police described this technical method as one they had not encountered before.

Source: NL Times

Banking & Lending

AI agents cut SBA compliance review times, according to a case study

First Internet Bank of Indiana has cut compliance review times for Small Business Administration (SBA) loan applicants by 87% using AI agents, according to a case study from Bretton AI. Reviews that took eight to 10 hours now take less than an hour.

−87%change in SBA compliance review times, per a Bretton AI case study

NoteThe case is about capacity, with the bank growing and more files for the same analysts. The $6 billion-asset digital bank plans to use the freed capacity to take on more business, according to PYMNTS.

Source: PYMNTS

Models & Labs

OpenAI's GPT-6.1 Terra has half the API price of GPT-5.6 Terra

OpenAI released GPT-6.1 Terra on Oct. 26, and its standard API prices are 50% lower than GPT-5.6 Terra and GPT-6.1 Sol, according to the release. Input tokens cost $1 per million and output tokens $5, compared to $2 and $10 for Claude Sonnet 5.5.

−50%standard API price change versus GPT-5.6 Terra, per OpenAI

NoteThe company says the model is practical for automating high-volume work such as reading invoices and entering their information into accounting software.

Sources: The Decoder, PYMNTS

Monday, 26 October

Banking & Lending

Itaú Unibanco has agreed to buy all shares of AI startup NeoSpace

Itaú Unibanco entered into a Share Purchase Agreement for all shares of NeoSpace, a Brazilian artificial intelligence startup developing models trained on Itaú's proprietary data. The purchase price will be BRL 2,925,000,000.00, with an installment payable at closing and the rest deferred over five years. Closing is subject to approval by the Brazilian Antitrust Authority (CADE) and the Central Bank of Brazil.

NoteItaú's filing says the deal represents another step in its strategy to enhance its proprietary artificial intelligence capabilities.

Source: Itaú Unibanco (SEC Form 6-K)

Saturday, 24 October

Banking & Lending

Customers Bank says an AI pilot closed select commercial loans in seven days

Customers Bank said a pilot of its AI loan process closed select commercial loans in seven days, against a typical 30 to 60 days. Its parent, Customers Bancorp, said on Oct. 22 that it is scaling the process and estimated more than $15 million in annualized capacity.

NoteEmployees said in a survey that they saved an estimated 76,000 cumulative hours, up nearly 65% from the second quarter.

Source: PYMNTS

Friday, 23 October

Banking & Lending

Goldman CEO Solomon says risk principles could be applied to AI oversight

Goldman Sachs CEO David Solomon wrote in a Financial Times article published Thursday that the principles Goldman uses to manage risk could be applied to AI oversight. They include holding individual decision makers accountable for the outcomes produced by their systems and routinely stress testing AI systems, Solomon wrote.

NoteSolomon wrote that the principles could be used by both the developers of AI and the companies that deploy the technology.

Source: PYMNTS

Thursday, 22 October

Top story: Banking & Lending

Major US banks are restricting external AI agent access, Axios reports

Several major US banks, including Morgan Stanley, JPMorgan, Bank of America and Wells Fargo, are restricting external AI agent access, according to Axios. Most want to limit agents to read-only access, letting them view account data but not carry out any actions.

NoteMany banks also can’t tell whether an agent is acting on a user’s behalf, and The Decoder says blocking agents on the provider side will be hard.

Source: The Decoder

Top story: Banking & Lending

Barclays plans to cut 5,000 jobs in an AI-linked efficiency push

Barclays plans to cut 5,000 jobs, about 5% of its workforce, as part of an artificial intelligence-linked efficiency push, CNBC reports. About 60% of the cuts are expected in support functions, including developers, according to the earnings presentation.

5,000jobs Barclays plans to cut, about 5% of its workforce

NoteBarclays tied the job plan to AI in its earnings presentation, and the cuts are expected to fall mainly in support functions. Other banks with similar functions may face the same question.

Source: CNBC

Wednesday, 21 October

Banking & Lending

Hackers reportedly used Claude and Grok agents in attacks that included banks

A hacking group known as CyberCartel reportedly used AI agents from several tech companies to automate cyberattacks on Latin American businesses. According to the report, the hackers used AI agents powered by models such as Claude and Grok to attack more than 400 targets in Latin America, including banks and payments companies. Anthropic said it quickly blocked the hackers’ use of Claude.

NoteStraiker’s CEO said enterprises are “just not ready”, and the report says many companies are now investing heavily in AI-powered cyberprotections.

Source: PYMNTS

Banking & Lending

Capital One card volume rose 11% while investing further in AI

Capital One's credit card purchase volume increased 11% year over year to $254.7 billion in the third quarter. Domestic card charge-offs fell to 3.98%, down 65 basis points from a year earlier. The bank is also investing further in generative and agentic AI.

11%year-over-year growth in credit card purchase volume, third quarter

NoteFor card lenders, falling charge-offs alongside higher spending is the number to watch next quarter.

Source: PYMNTS

Banking & Lending

Regional banks are putting AI into everyday banking operations

Third-quarter results from PNC, U.S. Bancorp, Truist, Fifth Third, Regions and KeyCorp show that regional banks are putting artificial intelligence into everyday banking operations. PNC reported average commercial loans increased $10.4 billion from the second quarter, and it attributed the increases to strong new production and higher utilization.

NotePNC's CEO described an "agentic workforce" that includes systems able to recognize AI agents as well as employees, so the operations work is more than a pilot.

Source: PYMNTS