The Broken Ladder: Why Simulation Has to Replace On-the-Job Practice
Edmundo Barrientos · 2026-10-28 · 3 min read
TL;DR: Because AI handles the routine tasks that once trained entry-level workers, structured practice must step in. However, to be effective, this practice must replicate the exact cognitive demands of real-world work.
Daniel's first hard conversation arrives on day one
Imagine Daniel, newly promoted to lead a team of six. Previously, new leaders began with small tasks, like running the weekly meeting, drafting status updates, or assigning routine work, each offering a manageable lesson in leading people.
Today, AI manages those routine duties, leaving new leaders to face hard people moments right away. Daniel's first is with a former peer whose project he must reassign. She says: "I thought we were friends."
Despite passing the company's leadership course, he freezes under pressure and reads out a line from the HR guidelines, leaving his colleague even more upset.
The bottom rungs have disappeared
Deloitte Insights identifies this challenge in research co-authored by Christine Elliott, Megan Cluver, and Joe Mariani, labeling it a broken skills ladder.
An examination of US job listings from 2022 through 2025 revealed a drop in entry-level data science hiring, even as senior-level demand remained robust. As the authors note, "The work that once taught people how to do a job is disappearing, even as the need for expertise persists."
To address this gap, Deloitte recommends options like mentorship, apprenticeships, and simulation-based training. Yet with no mentor available to sit in on every conversation, what form of initial practice should Daniel receive?
Matching the Cognitive Demands of the Job
Not all practice yields the same results, a key distinction missing from Deloitte's analysis.
While a standard quiz merely tests recognition of correct answers, handling a real moment like a reassigned project requires production. The leader must weigh the colleague's concerns, decide what to say, and say it in a way that keeps trust intact. Passive video content or multiple-choice assessments address the subject matter but bypass this vital thought process.
We call this necessary standard cognitive fidelity: ensuring practice directly replicates the mental processing required on the job, rather than just covering its topics. Ultimately, watching a good conversation never builds the judgment needed for real-world execution.
To restore this lost developmental step, we must rely on simulation. The AI models realistic responses based on input: staying guarded if the leader hides behind policy, or opening up when he acknowledges the real concern.
Because the system generates scenarios directly from the company's own past conversations, practice changes with each person, so each attempt presents a fresh dynamic. Team leaders retain full control over scenario design and readiness evaluations, exemplifying augmentation over automation.
Measuring Real-World Impact
Beyond individual scenarios, this problem is an enterprise-wide challenge. AI-driven simulations enable staff to navigate dozens of complex conversations safely, treating mistakes as learning points instead of costly damage with real colleagues. Authentic skill development isn't proven by finishing training modules; it shows up on the job. Meaningful skill transfer happens when a manager sees a new leader handle the next real hard conversation well, without help. Scaled across an entire organization, this practice becomes capability building: sustainable skill that shows up in the work. Conventional learning management systems and coursework successfully establish baseline understanding, but simulations offer the targeted arena needed to translate knowledge into everyday execution. The same ladder is breaking for sales, support, compliance and onboarding. Bridging the gaps in modern career progression demands structured practice aligned with real job complexity.
Related reading
- Autonomous Skill Building in the Flow of Work: Defining the New L&D Category
- Judgment Cannot Be Learned by Watching
- Augmentation Over Automation
References
- Deloitte Insights. AI is changing work. It might also be limiting how expertise is built, 2026. By Christine Elliott, Megan Cluver, Joe Mariani and co-authors.
About the publisher: this article is published by Grasply (grasply.ai), a capability builder. Grasply turns what a company already knows, its docs, calls and recordings, into skills people use at work, for employees, partners and customers. Practice turns knowing into doing, and Grasply shows proof that the work changed.