Who leads on AI

HSBC’s new Group Chief Executive Georges Elhedery is shaking up his leadership team. What caught our eye was a promotion: Group Chief Information Officer Stuart Riley, who came over from Citi last November, is taking on an expanded role.

The 2023 Evident AI Leadership Report found that HSBC was one of 14 banks (out of 50) without a CIO or CTO on their executive committee. This latest promotion is a signal of HSBC’s prioritization of AI in its strategy and increased AI maturity.

Now for more from the 2024 Leadership Report…

The market is clamoring for banks to explain their strategy amid rising costs (See “AI Costs Spiral,” The Brief, July 25). Our latest Leadership Report tells you how well they’re doing. The upshot: They’re beating the drum more often and in more places. The 50 Evident AI Index banks published over 1,250 references to AI across annual reports, press releases and company LinkedIn posts. That’s a 59% increase year-over-year.

BANKS GET SPECIFIC ABOUT USE CASES AND ROI

If all this talk didn’t include detail on use cases, measurable business impact and specifics on how banks plan to scale AI, then that’s just AI washing, plain and simple.

That’s not the case.

Last year, 27 of the 50 Index banks didn’t talk about their AI initiatives in detail in their investor relations materials. This year, the number of banks ignoring the elephant in the room fell to eight.

A NEW BEST-IN-CLASS?

Three banks stood out this year that had struggled on this score. CIBC, Charles Schwab and Groupe BPCE ranked 46th, 47th, and 48th, respectively, in the Leadership pillar in the 2023 Index. This year, all publicly shared their AI strategy, initiatives and outcomes in great detail. Expect the Leadership rankings to change when we release the 2024 Evident AI Index next month.

Join us on Tuesday, September 17 for our Evident AI Leadership Roundtable to discuss what the leading banks are doing and why to communicate their AI strategies.

Welcome to the debut of Meet the Model. In every Brief, we’ll highlight what’s new in LLMs that’s relevant to the banking industry.

Today: the latest 3.1 version of Llama.

What is it?

Meta’s flagship LLM comes in three sizes, from a beefy 405B model that competes with OpenAI and Anthropic’s best offerings, to a smaller 13B model that can handle simple queries quickly. (For more on model size, see “Bigger AI-n’t Better”, The Brief, July 11)

What does it do?

• You can run it on your own servers for free for up to 700 million active users before commercial licensing kicks in.
• Fluent in eight languages, including English, French and Hindi.
• Offered by virtually all cloud providers, including Google, Amazon and Microsoft.
• Comes with additional safety tools to guard against malicious requests.

Who uses it?

With 350 million downloads since 2023, Llama is the dominant open-weights model. Banks including Goldman Sachs, Citi and Nomura have said they use the model, and more are experimenting privately.

"If you’re only selling models, it's going to be difficult because it’s going to be like a zero margin business… it’ll still be a pretty high number because people need this tech – it’s growing very quickly – but the margins, at least now, are gonna be very, very tight."

- Aidan Gomez, CEO of Cohere, 20VC Podcast (August 19th)

In this week’s Use Case Corner, we look at three of the most interesting tools unveiled in the past couple of weeks.

And ICYMI: Check out our special edition Brief on 74 AI use cases.

#1 Gen AI-enabled IT Support

Use Case: ChatIT
Vendor: n/a
Bank: CommBank

Why it’s interesting: CommBank deployed a generative AI-enabled IT support chatbot to provide quick solutions to technology problems.

Potential ROI → increased employee productivity, increased employee satisfaction
Reported ROI → average response time of 14 seconds, over 10,000 employees have interacted and positively rated the experience

#2 AI-enabled Estate Management

Use Case: Estate Management Processes
Vendor: Pegasystems
Bank: Wells Fargo

Why it’s interesting: Wells Fargo embedded AI tools in its estate management process, following a six-year-long effort to integrate its various IT systems.

Potential ROI → improved efficiency, increased customer satisfaction
Reported ROI → NPS scores increased from “negative territory” to over 60 for bereaved families interacting with the bank from both IT integration and AI enablement

#3 Gen AI & Multi-LLM Productivity Tool

Use Case: CIBC AI
Vendor: n/a
Bank: CIBC

Why it’s interesting: CIBC has built a tool – CIBC AI -- with access to “multiple” large language models that automates regular, routine tasks, such as summarizing documents or drafting content. The pilot was launched in late July.

Potential ROI → increased employee productivity, increased employee satisfaction
Reported ROI → n/a

Have feedback on or ideas about use cases? Let us know at [email protected].

AI significantly reduced costs at fintech lender Klarna, CEO Sebastian Siemiatkowski said on their latest earnings call, letting the company reduce its headcount almost by half as it plans its IPO.

AI makes bad investment decisions, according to Scientific American. Fully and partly AI-driven investment funds on average brought lower returns than if you put your money on the S&P 500.

We are two or three years away from an AI financial advisor that meets SEC expectations, says Andrew Lo, professor of finance at MIT Sloan School of Management. The issue is trust. Existing chatbots are just role players, and their outputs are often too generic and glib; designing them to engage emotionally with users will be key, says Lo.

South Korea will relax regulations to boost AI adoption in the financial sector, vows the chief of the Korean Financial Services Commission. Regulators elsewhere will be watching to see what this means for innovation in finance.

Banks know about the risks AI poses to their business (see “Banks have a deepfake problem”, The Brief, August 22). Roles to manage these risks are typically few in number relative to the overall AI workforce. This makes the following three from August noteworthy.

Lloyds is looking for a new head of AI & data ethics to lead its strategy from within the newly formed AI center of excellence. The previous occupant, Paul Dongha, left for NatWest three months ago, Lloyds has set the bar high: You need a PhD in AI or AI ethics and several years of technical experience to apply.

ING has a vacancy for a senior GenAI model risk specialist. The focus is to ensure that Gen AI models comply with internal procedures and policies. We’ve not seen many of these specialized roles to date, but expect more.

BMO is hiring for a manager, AI risk to track how the company manages risks across its AI use cases, and provide regular reports to senior executives on new AI risks.