Traditional learning models struggle when roles, knowledge and context shift faster than training can keep up. Dr. Ravinder Tulsiani argues for a two-layer capability model combining structured preparation with real-time support, helping organisations strengthen decision-making, reduce costly errors and position L&D for improved performance rather than content delivery and completion.
Learning and development was not designed for continuous, high-velocity change. It was built for environments where roles were stable, knowledge had a longer shelf life and change was episodic, not constant. In that world, the model made sense:
Train people → deploy them → refresh periodically
Today, that sequence breaks because roles are shifting in real time, decisions are more complex and artificial intelligence is accelerating both the pace and the consequences of those decisions. Yet many L&D functions are still operating as if the problem is content coverage. It isn’t. The problem is decision volatility.
The real shift: From knowledge delivery to decision stability
The question is no longer “Did people complete the training?” but “Can they make effective decisions under changing conditions?”. Most learning models were designed to transfer knowledge ahead of time, but in high-velocity environments, two things happen:
- Knowledge decays before it is used
- Context changes before it is applied
This creates a false sense of readiness. People feel prepared because they have seen the material, but when the moment comes, the situation is different. This is where traditional L&D breaks, not because it is wrong, but because it is incomplete.
The overcorrection that makes things worse
In response to these issues, many organisations swing to the opposite extreme, moving everything into the flow of work, replacing training with performance support and using AI to guide decisions in real time. This sounds modern, but it’s also dangerous.
Learning should not happen for the first time in a live environment. No one wants a surgeon reviewing a procedure for the first time while a patient is on the table and the same principle applies across domains. Some capabilities require pre-loaded competence, whilst others can be supported dynamically. The failure is not in either approach, it’s in treating them as interchangeable.
The model that actually works: A two-layer capability system
High-velocity environments do not eliminate the need for structured learning, they demand better placement of it. The solution is to separate capability into two distinct layers: foundation and flow. Each serves a different purpose and is triggered by different conditions.
Foundation: What must be learned before action
The foundation layer exists for one reason: To ensure that people are safe and competent before they are exposed to real consequences.
This applies when:
- The cost of failure is high
- Errors are irreversible
- Decisions carry legal, financial, or safety risk
In these cases, learning cannot be deferred, it must happen in a controlled environment. This is where simulation, rehearsal, and structured practice matter, not content for the sake of coverage, but exposure to realistic scenarios where mistakes can occur without real-world damage.
The goal is not completion, but a clear threshold of whether the person is ready to perform, or not. If that threshold is not met, nothing else matters.
Flow: What must be supported during action
The flow layer operates in a different reality.
This is where:
- Conditions change rapidly
- Edge cases emerge
- Judgment is required under pressure
Here, the issue is not initial competence, it is consistency under variation and it’s where
support becomes critical. This isn’t about teaching for the first time, but using those materials to stabilise performance.
Effective flow support does three things:
- Clarifies when a situation requires attention
- Guides decision pathways
- Highlights common failure patterns
Artificial intelligence becomes powerful here, not to generate content, but as an adaptation layer as it can detect patterns, update guidance, and surface relevant signals in real time. It’s reinforcing capability, not replacing it.
The critical discipline: Decision classification
The credibility of this model depends on one capability: Knowing which decisions belong in which layer. Every learning effort should start with four questions:
- What decision is being made?
- What is the cost of failure?
- Is failure recoverable or irreversible?
- How stable is the environment?
From there, classification becomes clear:
- High-risk, irreversible decisions → Foundation-heavy
- Moderate-risk, variable decisions → Hybrid
- Low-risk, rapidly changing decisions → Flow-heavy
Without this discipline, organisations either overtrain everything or underprepare what actually matters. Either one will destroy credibility.
Why this changes the role of L&D
This is not a shift in delivery methods; it’s a shift in responsibility. L&D is no longer responsible for producing learning.
It is responsible for ensuring that critical decisions are performed correctly, capability is established before exposure where required and performance is stabilised under real conditions.
This changes the conversation with the business from “How many people completed the programme?” to “Where are decisions breaking, and how are we fixing them?” And it’s a fundamentally different value proposition.
What this looks like in practice
Instead of designing courses, L&D identifies moments that matter, where errors are costly, variation is increasing and performance is inconsistent. For each moment L&D need to define the required level of pre-performance competence, design targeted simulations (if needed), and embed support for real-time decision-making.
Then measure what actually matters:
- Are decisions improving?
- Are errors decreasing?
- Is confidence aligned with capability?
If the answer is no, the system adjusts. Continuously.
The bottom line
High-velocity environments do not require more learning, they require better alignment between when learning happens, how it is applied, and what decisions it is meant to influence. The goal is not to eliminate training, it’s to place it precisely where it belongs, and reinforce it where it is most needed.
Because in the end, organisations are not competing on what people know, they are competing on how well people decide when it matters.
Dr. Ravinder Tulsiani is Principal, Interim Enterprise Capability & L&D Advisor at Catalyst Learning
If this article piqued your interest, you’ll probably find value in our conference!
Find out about keynote speaker Erica Farmer, her opening session on the
AI Fire Hydrant and what it means for productivity, performance and L&D
Don’t miss 50% off your ticket for the first 30 only!
