America’s biggest bank wants every member of its more than 300,000-strong workforce to be an expert on how to juice AI.
The bank, with its $18 billion technology budget, has invested extensively in AI development. Now, it’s turning its attention to a firmwide training push. The goal? To educate its worldwide workforce on how to make AI work for every single employee — not a one-size-fits-all approach, Derek Waldron, chief analytics officer at JPMorgan Chase, said in an interview with McKinsey about the bank’s AI adoption strategy.
“Training needs are varied, just like AI applications. The best way to approach this is segment by segment,” Waldron said in the interview, which was published on the consulting firm’s website this week. Everyone from rank-and-file workers to company leaders will have to learn new skills, Waldron continued.
That being said, JPMorgan launched an internal training program for beginners, “AI Made Easy,” he said, adding that “tens of thousands” had already taken the course. The firm has created modules to educate users on how to conduct thorough research with AI or get the most out of several different data sets.
It’s not just the managed who may need to change their ways — it’s the managers, too. Waldron predicted that CEOs and business leaders will have to adopt new approaches as the tech’s reach becomes more widespread. “Value from gen AI won’t come just from giving people tools; business leaders must lead cross-functional teams through transformation in the age of AI,” he said.
A multi-pronged strategy — from town halls to communications from managers to marketing campaigns on screens across the bank’s offices — is helping to get people comfortable, he said.
What training looks like
Teaching people about AI comes down to two main layers, Waldron explained. Step one: What can AI large language models do and not do? And step two: How do you formulate the right questions?
“Once there’s familiarity with capabilities,” he said, “we move into how to construct good prompts, with frameworks and examples and constraints.” Then, things get more sophisticated: “how to pivot the persona of an LLM from maker to checker, or how to use two LLMs to debate a concept to get more creative.”
This transition is a companywide endeavor, with workers often teaching one another.
“Many teams quickly set up prompt libraries, ‘prompt of the week’ emails, and social channels to share power-user innovations,” he said, adding: “If we get the technology into employees’ hands — with change management and training — they’ll be best positioned to innovate and put it to good use.”
Everyone will have to make changes
Waldron also offered insight into how some technical roles are learning new skills.
“Software engineers need to be upskilled to build scalable AI systems based on agents and LLM components,” he said, adding: “Another population is technologists, who will increasingly want to build sophisticated applications using agentic or gen AI. That skill set is something that needs to be trained.”
Agentic AI is drawing significant focus from Wall Street and Silicon Valley. The notion that semi-autonomous digital “agents” could orchestrate end-to-end projects independently has been polarizing in some quarters. But at a conference last week, Teresa Heitsenrether — JPMorgan’s chief data and analytics officer, to whom Waldron reports — said she thought managing battalions of digital agents would give early-career workers a taste of being a boss earlier than they might get otherwise.
For data scientists, Waldron said that tech advances meant the days of building standard models were over. Third-party providers tend to handle this now, he said, letting in-house data scientists evaluate and enhance ready-made models and “apply their skills to designing, evaluating, and optimizing systems.”
In other words — the fun stuff.
Read the full article here
