New: 2024 Talent Report

The 50 banks we track for the Evident AI Index added 9,238 net new AI roles and posted 17,836 new AI-related job listings since our previous report in November of last year.

Find out how leading banks are thinking about AI talent and explore the “banks to watch” in the 2024 Talent Report, published today. The full report is available to Evident members. Non-members can download an excerpt. To learn more about membership, please contact us.

"The spread of the application of AI by the private sector and its encouragement by appropriate government policy is the only answer to Britain’s productivity challenge and, over time, it can turbocharge growth."

- Former British Prime Minister Tony Blair,
writing in Sunday Times, July 6

Keir Starmer will be looking to AI and the City of London to boost the new Labour government’s plans for British revival. Those would seem to be two of the U.K.’s greatest built-in advantages.

But the smart conversation in London is about how the place may be losing its mojo. While some top-tier AI start-ups like DeepMind and ARM were born here, and in Europe the U.K. attracts a large share of VC activity, the center of AI innovation has clearly moved eight time zones away to the Pacific coast of the U.S. That’s also where capital is increasingly concentrated.

It doesn’t mean there isn’t a good market for talent here. Microsoft and OpenAI opened European hubs in London this year, and Google bought DeepMind more than a decade ago. Banks are hiring fast too. In the last six months, the number of AI development roles at the 50 Evident AI Index banks has grown faster in London – up 6.1% – than other major hubs like New York (+3.7%) and Toronto (+5.9%) (see our latest Talent Report). But most of those jobs are at non-British companies. In the City, AI developers are twice as likely to be working for foreign banks as British ones.

That’s not to say British banks aren’t driving toward AI maturity. We’ve heard a lot from the CEOs of Lloyds and HSBC about AI in the last few months. At last month’s Evident Symposium in London, Zachery Anderson of NatWest spoke about how they’ve improved the speed at which they can find value from AI use cases. And Barclays recently created a Center of Excellence for generative AI, as Kristen Bennie explained.

But the fact is that Britain is more like an AI outsourcing center for the tech giants and American financial institutions. It’s hard to claim you’re the leading nation on AI when the actual leaders – both the big players and the most interesting start-ups – are somewhere else.

BIGGER AI-N’T BETTER

Speaking of would-be European AI hubs, in Paris yesterday the AI upstart Mistral inked a deal with BNP Paribas to support the bank’s customer support, sales and IT operations.

Expect to see more deals like this. As the costs of using AI rise faster than the revenue generated by it – as widely-shared reports from Sequoia and Goldman Sachs recently pointed out – banks are looking to smaller and cheaper models.

Meta’s Llama model tops out at 70 billion parameters – the values that set how the model responds to questions – compared to OpenAI’s GPT-4 (1.8 trillion). But Llama is 33 times cheaper to run. That’s why Wells Fargo, Citi and Goldman Sachs all use it.

Join us and our distinguished panel of guests next Tuesday for our 2024 Talent Roundtable: How can banks close the AI talent gap?

We’ll dive into the latest data from Evident’s 2024 Talent Report and discuss: What skills do banks need in order to prepare for – and thrive in – an AI-first future? How are they going about attracting, developing and nurturing this talent? And what role can external partners play in building capability in the short- and long-term?

#1 Email assistant

Use Case: Debrief
Vendor: OpenAI
Bank: Morgan Stanley

What it is: As part of Morgan Stanley’s ongoing partnership with OpenAI, the bank has launched Debrief, a meeting assistant to support financial advisors. With a client’s permission, the tool joins in meetings, and records, transcribes and summarizes them, and can automatically draft emails based on discussions.

Why it’s interesting: As the bank plans to release the program to its 15,000 advisors by early July, this is an example of the rise of “Agentic AI.” (See “Coming to a bank near you: Agentic AI… and Quantum,” The Brief, June 22).

Potential ROI → increased employee efficiency, increased accuracy, increased customer satisfaction
Reported ROI → one employee reported the tool saved them about half an hour per meeting

#2 Fraud and mistake prevention in payments

Use Case: NameCheck x Liink by J.P. Morgan
Bank: CommBank x JPMC

What it is: Two AI heavyweights have teamed up to tackle payment fraud and errors in payments. CommBank’s AI-based NameCheck security tool reviews account details to determine whether entered account details are correct, and has been piloted on JPMC’s B2B Liink payment network.

Why it’s interesting: Public inter-bank partnerships on AI use cases are rare, but not unheard of. CommBank itself has a history of making its AI models available for other banks (see its abuse detection model), while HSBC and Google’s anti-money laundering AI product is available for other institutions. Tackling the big issues, like financial crime, may need an open approach from banks, working together to create the best AI use cases.

Potential ROI → reduced fraud, increased payment accuracy
Reported ROI → n/a

#3 Data restructuring

Use Case: Restructuring Open Banking Data
Vendor: n/a
Bank: Mastercard

What it is: Mastercard manages a huge amount of unstructured data. On a recent podcast, Jess Turner, executive vice president of global banking and API at Mastercard, confirmed the bank is using AI to make its messy data usable.

Why it’s interesting: We keep hearing how banks can use AI to get more from their data. “I think what organizations will quickly start doing is thinking about integrating AI strategically across the data lifecycle – and that’s where you'll see a convergence of the data and the AI strategies as the hype dissipates,” Rinesh Patel, global head of financial services at Snowflake, said at Evident’s London Symposium.

Potential ROI → shorter time to production of AI use cases
Reported ROI → n/a

CommBank announced yesterday that their venture-scaling arm is offering a $250,000 investment and a pathway to partnership for an early stage data & AI start-up to reimagine the customer and employee experience. A good example of how a venture program is being aligned with a bank’s AI strategy.

The Basel Committee of banking regulators proposed that boards of banks should shoulder ultimate responsibility for oversight of outsourced services from cloud providers. The committee wants boards to document how they make decisions about third-party digital services, which have increasingly come under scrutiny because of heightened susceptibility to cyber hacks.

Bain’s quarterly survey on AI finds that while last year, “executives were most concerned about quality and capabilities… in 2024, many are already focused on delivering real value.” Sales, coding, and marketing use cases exceeded expectations. Legal and HR implementations have been underwhelming. Six in 10 of companies see AI as a top priority, but only 35% say they have a clear value-generating plan.

JPMorgan is paying up to $325,000 for quantum talent, hiring 28 PhDs across New York and Singapore. The bank is pursuing quantum with an eye to AI, using it to accelerate training for AI models.

A lot of data at Citi is still in Excel spreadsheets and not ready for the bank’s ambitious AI rollout. CIO Shadman Zafar’s attempt to clean up 6,000 legacy systems has been slowed by unusable data in manual systems or spreadsheets. To many executives’ chagrin, it’s not enough to simply implement LLMs – companies need to sort out their underlying data.

Technological impact is “not a law of nature… it is by no means guaranteed” says MIT economist Darren Acemoglu in a cautionary Goldman Sachs report on generative AI spending. (See “How Goldman Drives AI Conversation, The Brief, June 27).

Citi’s Head of Risk Data Peter Cai left the bank, as part of an ongoing restructuring. The move comes months after CEO Jane Fraser said on an earnings call that Citi is “intensifying its efforts to automate regulatory processes and data related to regulatory reporting.”

AccessFintech hired Naveen TV to lead its product group. He comes from JPMorgan, where he was global head of data strategy and AI.

Andrew Mulligan joined Westpac as international head of technology. Mulligan was previously at AWS where he was head of technology for their Singapore branch.

Know an event that you think should be featured? Let us know at [email protected].