Special Edition: Execute. Enable. Innovate.

Over eight hours of conversations on stage, and more in the corridors and dining tables outside, some themes came up repeatedly. Here are the three most eye-catching ones:

#1 Execute!

If 2023 was the year of experimentation, Nvidia’s Jochen Papenbrock told the audience, 2024 is all about industrialization.

Far more use cases are moving down the “AI Factory” line. On Wednesday, we heard numerous updates on sheer numbers: Deutsche Bank has 200 use cases in the pipeline, some going into production this year. Natwest: 400, 100 “prioritized”. DBS:  350. BNY: 700 (more or less) use cases. They were reluctant, however, to discuss specifics on which ones are most promising before any formal rollouts.

And they are able to get them out the door faster. Zachery Anderson at NatWest said the time from proof of concept (POC) to production came down from “24 months on average to something more like 90 days” since he joined the bank in 2020.

Speed matters. AI teams can decide if the POC is likely to generate value earlier, and then redeploy resources or cherry pick what they’ve learned or liked to other projects. “Reusability” is a new buzz word we learned yesterday.

#2 AI must deliver for consumers and bottom line – not for its own sake

There’s growing pressure from the markets and boards to show the value from the huge investment in this emerging technology.

So another priority now is to figure out how to “enable” and scale a bank’s AI strategy by building strong guardrails, fine tuning data strategies and putting a lot of effort into training people across departments on how to use new AI applications.

JPMC’s Guy Halamish said his bank uses the following framework to assess use cases and whether to move ahead with them:

1. Opportunity: What’s the size of it (e.g. efficiency gains, customer engagement etc.)?
2. Capability: Are there existing capabilities in the bank to put the use case into production?
3. Motivation: And is there a leader or team that’s energized enough by it to make it a success?

Offering a different approach, Dan Jermyn at CommBank preached holism.

“We have tilted away from looking at a straight ‘use case by line of business’ approach, towards building capability and enablement,” he said. “As it becomes easier to deploy AI, there is less of a need to balance and compare the ROI of this use case versus others – and move towards everyone across the bank being able to provide a solution after they’ve spotted something that can make a customer's life better. And you can't do that easily if you have a complex sliding scale of what does or does not get built.”

#3 Coming to a bank near you soon: Agentic AI…and Quantum

Everyone keeps saying humans must stay in the loop, but that tune changes when people look beyond the horizon. There they see – with conviction if not quite full clarity – that AI Agents will be a lot more autonomous and that quantum computing is the next transformative technology.

If intelligence might be defined as perception, cognition and action, current AI applications are only doing the first, noted Manuela Veloso, head of AI research at JPMC. This is where agentic AI comes in, which would see the models act on their own. Her example is an AI Agent that will fully answer the piles of daily emails in our inbox.

“If you believe in an AI future, you have to build it first,” said Tony Kim, head of tech sector fundamental equities at BlackRock. Over the last two years, we’ve built the foundation with computing power. To have a robust AI ecosystem, you have to layer on cloud, model, data, applications and services, he said.

As for quantum computing, which would let us solve problems that current systems can’t, that’s “as close to the future as possible,” said Chris Bishop, the host of Inside Quantum Technology podcast, getting a laugh from the audience.

That may be sooner than we realize. JPMC built an 80-person strong team to work on quantum and invested $100 million in Quantinuum. The company’s co-founder, Ilyas Khan, told us yesterday that quantum computing would arrive by 2027. Possible use cases? Derivative pricing, portfolio optimization and stochastic modeling (translation: accounting for randomness in insurance and financial models).

You can watch each panel session here.

Join us for our 2024 Talent Roundtable: How can banks close the AI talent gap?

In this roundtable, we'll dive into the latest data from Evident’s upcoming 2024 Talent Report to consider: What sorts of 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?

Throughout the Symposium, our AI transcription tool kept a tally of the top “buzz words” mentioned on stage. A scoop: AI didn’t even come in first. Data did.

Here’s the leaderboard.

PS: Why Sheryl Crow? Well the American singer recently released ‘Evolution’, where where she shares her feelings about AI: “Where are we headed in this paradise?” she sings. “We are passengers and there's no one at the wheel. No matter how well you can outdo me. There is one thing you will never do and it's feel.”