The currency is being devalued in real time
For as long as any of us have been working, professional standing in financial services has rested on a simple equation. Know more than the person next to you. Retrieve it faster. Apply it under pressure. That is what a graduate program selects for, what a promotion committee rewards, and what makes someone the person others call before a client meeting.
Generative AI is devaluing that equation, and it is not doing it slowly.
Cameron Hedrick, founder of Hedrick Advisory and former Chief Learning and Culture Officer at Citi, framed the shift this way in conversation with Sarah Clarke, CEO of Intuition Americas: the premium on wisdom and discernment is going up, precisely because those are the things machines do not do well and people can be extraordinary at. His warning to anyone sitting in a CHRO, CLO, or board seat was less comfortable. If you understand even a little about how wisdom is actually acquired, he said, you should be terrified.
He was using hyperbole to make a point land. The point is still correct.
Wisdom is knowing what matters
Wisdom is a word that gets used loosely enough to mean nothing, so it is worth pinning down. Cameron's working definition is four words: knowing what matters. Not what you can do. What you should.
Sarah offered the sharpest illustration of it, from her own background in design. Photoshop has two things inside it. There is the technical layer, knowing how to grade the color, mask the layer, run the adjustment across a thousand files instead of one. And there is the other thing: deciding whether the adjustment serves the picture at all.
Cameron took the analogy where it needed to go. AI explodes the number of things you can do. It says nothing about which of them you should. "The human has to put his or her hand on it to mold it and shape it and make it something of value."
In a bank, that distinction is not aesthetic. It is the difference between a memo that is fluent and a memo that is right. Between a credit narrative that reads well and one that survives challenge. Between a model output somebody accepted and a model output somebody tested. Fluency now comes cheap. Judgment about fluency does not.
The capabilities rising in value are the awkward ones: ethical judgment, discernment, the ability to create conditions where a team will actually say what it thinks, systems thinking, and the political and contextual reading of a room that no model can map. Cameron added one more that deserves attention in a regulated environment. Most of us think in binaries. The work ahead demands comfort with paradox and with situations that have no clean answer, which is exactly the shape of most real risk decisions.
The part that should worry L&D
Here is the problem. Wisdom does not respond to the delivery model our function was built on.
"You can't microwave it," Cameron said. "It's a crock pot or a slow cooker. It's got to be there for a long time."
It accumulates through consequential decisions made in ambiguous conditions, followed by reflection, usually after something went wrong. It builds over years. You cannot assign it a completion percentage, compress it into a learning journey, or evidence it in a post-course assessment. Every instinct the function has been trained on says close the gap with more content. That instinct will not work here, and adding modules will look like progress for about two quarters.
There is a related problem underneath it that nobody enjoys discussing. Sarah named it directly: if the visible effort disappears, if the hard part of the work is no longer seen, what is a person bringing to the table? That question is not abstract to a VP who has spent fifteen years being the one who does the difficult analysis. It is the emotional weather your program will be delivered into, and pretending it is not there does not help anybody move through it.
What you can actually build
Three things, in order of how quickly they can start.
Real permission to experiment, with room to fail. Not a curated academy with a certificate at the end. If your institution is still gatekeeping who is allowed to try, capability is already accruing to whoever ignored the memo.
Structured reflection inside everyday work. Ask why that choice, not just what was produced. Ask what the implications look like across one month and across three years. Ask what got verified and what got taken on trust. Practiced consistently, that line of questioning builds the cognitive habit that judgment runs on, and it costs nothing but a manager's attention.
Simulation over instruction, wherever you can afford it. This is where Cameron would put his own program budget. Put people in front of ambiguity and ethical situations that have no clear answer, rather than more binary technical training. For financial services that is not hard to source. Your risk register is full of scenarios with no clean answer, and most of them never make it into a learning design.
Underneath all three sits the thing Cameron called the golden key: self-knowing. Being able to ask, honestly, why do I think this is the right call. Where is my ego in this. Is this intuition or is it analysis, and can I tell the difference right now. Some people do this naturally. Most have to be taught, and it takes a great deal more doing than a workshop.
What technology does not do
Sarah closed with a company rather than a framework, and it is the most useful thing in the conversation.
Rathborne's has been making candles in Ireland since the seventeenth century. Their candles once lit the streets of Dublin. Electric light should have ended them. It did not. They sell more candles today than they did then, because the candle stopped being a necessity and became a luxury.
Technology rarely deletes the thing. It changes what the thing means, and the change in meaning creates work nobody had thought of yet.
Which is roughly the argument for developing the human side of your organization now rather than after the org chart moves. Cameron's closing thought was about what stays durable. As long as people exist there is something specific to us. We feel pain. We can love someone or fail to. We hold perspective across disciplines and across decades in a way no model does. These are not the easy skills. They are the hardest ones, and they are the ones worth doubling down on.
The wisdom economy is not a forecast. It is the operating environment. The question is whether your function is building for it or still shipping content into it.