The first draft was the classroom.
In credit, risk, operations and advisory teams, the analyst's first pass was where judgment formed: what to question, what to escalate, what an answer left out. AI now produces much of that pass, quickly and fluently.
The Conference Board's July 2026 research found that AI is eliminating the corporate apprenticeship model, and that talent shortages and skills gaps, not technology, are becoming the greatest barriers to realizing AI's full business value.
Oversight of AI is a skill, and skills need reps.
A reviewer who accepts a fluent answer has not reviewed it. Challenging AI output means asking why, checking the source and noticing what is missing, then owning the call. These are behaviors, and behaviors fade without frequent practice.
RCM ThinkLabs does not replace compliance training or controls. It develops the judgment those controls depend on.
Real decisions, with incomplete information.
In RCM ThinkLabs, people work through an immersive narrative. Characters have their own interests and hold things back. Information is incomplete, and there is no script to repeat, so people use the skill.
The measurement is never shown, so they decide instead of performing. Each session takes up to fifteen minutes and ends with private feedback.
The analyst gets the practice. The manager gets the opening.
Early-career staff get a record of their own growth from day one. Their manager gets an opening for each one-on-one through RCM Advisor, so coaching starts from what the person did.
The design intent is that managers grow into better coaches as the loop turns. We have not measured that.
Developmental by design.
Regulated firms are right to ask how any AI tool touches decisions about people. RCM ThinkLabs is built for development. It compares each person with their own opening baseline, and it does not rank people against each other.
It runs in the browser, and no integration is required. See security and deployment.
What changed for participants.
Production data from Wingbrace, an advanced engineering team, over 39 days, measured against each person's own opening baseline. These results are from an engineering team, not a financial services firm.
A cohort, a baseline, and thirty business days.
A cycle starts with about thirty business days for one cohort, which gives each person a measured baseline and first movement. Lasting change takes repetition beyond that, and it depends on participation, feedback and managers reinforcing it.
Questions from financial services leaders
How do financial services firms develop early-career judgment when AI does the first draft?
Replace the lost repetitions on purpose. Analysts need frequent decisions with incomplete information, consequences and feedback, plus managers who coach from what each person did. RCM ThinkLabs provides the practice in short recurring sessions and gives managers an opening for each one-on-one.
Is RCM ThinkLabs a compliance or AML training tool?
No. It does not replace compliance training, controls or monitoring. It develops the judgment those controls depend on: asking why, checking the source and owning the call. It is a development tool, and it does not rank people against each other.
How does RCM ThinkLabs fit a regulated employer?
It is built for development, not evaluation. Each person is compared with their own opening baseline, the measurement is never shown during a session, and it runs in the browser with no integration required. Security and deployment detail is on the security page and available in a briefing.
Can it work alongside an existing graduate or analyst program?
Yes. A cycle sits beside the firm's existing early-career structure, so each class starts with a measured baseline. It adds short recurring practice and a coaching opening for managers. It does not replace rotations, training or mentoring.
How long before a manager sees change?
A cycle starts with about thirty business days of short sessions, which gives each person a measured baseline and first movement. Lasting change takes repetition beyond that, and it depends on people taking part, feedback landing and managers reinforcing it.
Where do the results come from?
From Wingbrace, an advanced engineering team, over 39 days: 84% of regular participants improved on a full read, and the share of questions that asked why rose from 1 in 17 to 1 in 7. They are not financial services results, and lasting change depends on participation, feedback and manager reinforcement.
Sources. The Conference Board, Human Capital Center, "Reinventing the Talent Pipeline in the AI Era", 10 July 2026. Deloitte, 2026 Gen Z and Millennial Survey, 13 May 2026, 22,595 respondents in 44 countries. Wingbrace figures are RCM ThinkLabs production data from an advanced engineering team, Days 1 to 12 against Days 28 to 39, measured against each person's own opening baseline. If anything here is out of date, tell us and we will correct it.
Put the first-draft reps back. Coach from what people did.
Comparing options? See how RCM ThinkLabs compares with Workera, BetterUp and other alternatives.