Financial Services Is Not Training for AI Tools. It Is Training for Judgment.

Look Past the Enrollment Numbers: What Financial Firms Actually Measure

There is a fast way to learn what an industry believes about AI readiness, and it is not reading its strategy decks. Look instead at what it is putting its people through.

Across leading financial services institutions, more than 2,500 employees are now enrolled in Workera-backed programs spanning 12 distinct tracks. The list of what those employees are being assessed on turns out to be more revealing than the enrollment figure.

What the sector is actually measuring

The most widely assessed capabilities are AI Essentials, Generative AI Essentials, and AI Ethics. Close behind sit Beyond LLMs: Prompts, Agents, and RAG, along with Project Management, Machine Learning Fundamentals, and Business Process Reengineering. Further down, Effective Communication Essentials and Problem Solving Essentials round out the curriculum.

Two things stand out.

First, AI Ethics ranks in the top three. Not as a compliance module bolted onto the end of a technical track, but as one of the capabilities institutions most want measured. Second, the tail of the list is not technical at all. Communication and problem solving are sitting alongside machine learning in the same portfolio of programs.

That combination is not an accident of catalog design. It describes a sector that has decided AI fluency without judgment is a liability rather than an asset.

Why judgment is the differentiator here

The 2026 Global AI in Financial Services Report from the Cambridge Centre for Alternative Finance, produced with the Bank for International Settlements, the IMF, and the World Economic Forum, found that three-quarters of firms saw AI change internal demand for skills during 2025.

Every industry is absorbing that shift. Financial services absorbs it under conditions most do not face. A model that produces a plausible answer in a marketing function produces a bad quarter. The same model in credit, in surveillance, in advisory, in disclosure produces a regulatory finding.

"In financial services, being AI-ready is not only about using new tools. It is about using them with the judgment and rigor the industry demands," said Kian Katanforoosh, founder and CEO of Workera. "These institutions are not guessing at readiness. They are measuring it, building against it, and proving the gains."

The breadth of programs is what makes that possible. Rather than pushing one AI literacy course across an entire population, institutions are directing each role toward the capabilities that role genuinely requires. A quantitative analyst and a relationship manager need different things from the same technology, and treating them identically wastes the budget on one and underserves the other.

The number that separates enrollment from progress

Participation is easy to report and easy to overvalue. Seat time has never been evidence of capability, which is why 30% of L&D budgets go to activity rather than skill gain, according to McKinsey.

Among employees who have been reassessed so far, average proficiency rose by 79 points on a 300-point scale. That gain holds across both AI-specific and broader data capabilities.

The mechanism behind that figure matters as much as the figure. A baseline tells leaders where the workforce stands on the day it is measured. Only a second measurement, taken after development, answers the question executives are actually asking, which is whether the investment changed anything. Most organizations never take that second measurement, so the honest answer to a board asking about return is that nobody knows.

"When average scores climb nearly 80 points on a verified scale, you can see upskilling turning into real capability, not just attendance," Katanforoosh said.

What this looks like inside an institution

Nothing about the pattern above requires a specialized function or a multi-year program design. It requires three things in sequence.

• Baseline what people can do, measured directly rather than inferred from resumes, self-ratings, or course completions.
• Direct development toward verified gaps, so budget flows to the capabilities each role is short on instead of spreading evenly across a population.
• Reassess, so the gain is a number leaders can defend rather than a claim they have to trust.

Institutions running that loop are not making a case for AI readiness. They are producing evidence of it, which is a materially different position to be in when a regulator, a board, or a chief risk officer asks.

The question worth sitting with

If someone asked today which of your teams can be trusted to apply AI to a regulated decision, where would the answer come from?

If it traces back to a completion report or a self-assessment survey, you have a proxy standing in for proof. There is a better starting point available.

Financial services leaders ready to baseline, build, and verify AI readiness across their workforce can request a demo.

‍