Training Happens in a Portal. Work Happens Everywhere Else.
There is a well-documented gap in corporate learning that almost every organisation recognises but very few have managed to close. Employees receive training in one environment, a learning portal, a classroom, a scheduled webinar, and then return to their actual work, where the conditions, pressures, and contexts rarely match what the training prepared them for.
The transfer problem has been studied extensively. Research from the Association for Talent Development and other industry bodies consistently shows that a significant portion of what employees learn in formal training never gets applied on the job. The reasons are predictable: the content was too general, the timing was wrong, the context was missing, or the gap between learning and application was too long.
This is not a content quality problem. Most organisations produce competent training materials. It is a proximity problem. The learning happens in one place and the performance happens in another, and the distance between them eats most of the value.
In-workflow learning agents close that distance by putting development where the work actually happens.
What an In-Workflow Agent Looks Like in Practice
The clearest way to understand an in-workflow learning agent is to picture an employee working through a normal business process.
Say a procurement manager is reviewing a new supplier contract. In a traditional model, whatever training they received on supplier risk assessment happened weeks or months ago. They may or may not remember the key frameworks. They may or may not have the latest policy updates. They are relying on memory and habit.
With an in-workflow agent, the picture changes. As the manager works through the contract review in their procurement platform, an agent sidebar provides contextual support. It surfaces the organisation's current risk assessment criteria. It flags clauses that warrant closer review based on recent policy changes. It offers a short refresher on a negotiation technique that is relevant to the specific type of contract they are reviewing.
The development is happening inside the task, not in a separate portal the employee needs to navigate to. The timing is immediate, not scheduled for next quarter. The content is specific to what the employee is doing right now, not generalised for a cohort.
This pattern applies across functions. A sales representative preparing a proposal receives coaching prompts based on the specific deal context. A project manager updating a risk register sees guidance drawn from the organisation's best practices and recent lessons learned. A customer service agent handling an escalation gets real-time support that reflects both company policy and the specific customer history.
In each case, the agent is embedded in the workflow tool the employee already uses. There is no context switch. No separate login. No interruption to the task at hand.
Why Proximity Changes Everything
The core insight behind in-workflow learning is that proximity to the task dramatically increases the likelihood that learning will transfer into practice.
When an employee receives a coaching prompt while they are actively working on the relevant task, the cognitive load of transfer drops to nearly zero. They do not need to remember a framework from a training session three months ago. The framework appears in front of them, adapted to their current situation, at the moment they need it.
Josh Bersin's research on learning in the flow of work has consistently shown that employees overwhelmingly prefer learning that is integrated into their daily activities over formal training programmes. The preference reflects something deeper than convenience. People learn more effectively when the content is immediately relevant and immediately applicable.
For organisations, this has a direct impact on the return on learning investment. If formal training produces capability that decays before it gets applied, the effective cost per applied skill is much higher than the training budget suggests. In-workflow agents reduce that decay because the gap between learning and application essentially disappears.
What This Means for L&D Teams
The shift to in-workflow learning does not eliminate the need for formal training. Compliance programmes, foundational skills development, and structured leadership curricula still have their place. What changes is that formal training is no longer expected to carry the entire weight of workforce development.
L&D teams that adopt in-workflow agents find that their role evolves in a few important ways.
First, they spend less time building standalone courses and more time curating the knowledge assets that agents draw from. The content library shifts from being a catalogue of courses to being a reservoir of modular content that agents can surface in context.
Second, they gain visibility into where learning is actually happening and how it connects to performance. In-workflow agents generate data on which prompts employees engage with, which they skip, and which lead to measurable changes in task quality. That data is far richer than course completion metrics.
Third, they become closer partners with operational leaders. When learning is embedded in workflows, L&D needs to understand those workflows in detail, which means deeper collaboration with the business units they support.
What This Means for Operations Leaders
For Heads of Operations and Transformation, in-workflow learning agents address a frustration that has persisted for years: investing in employee training and seeing inconsistent results.
The inconsistency is structural. When training and work are separated, the quality of transfer depends on individual motivation, memory, and opportunity. Some employees apply what they learned immediately. Others do not. The result is uneven capability across teams, even when everyone completed the same programme.
In-workflow agents reduce that variance by standardising the availability of learning support at the point of work. Every employee working through a particular process has access to the same contextual guidance. The floor rises, which means operational consistency improves.
For organisations in the middle of transformation programmes, where new processes, tools, and ways of working are being introduced at pace, this is particularly valuable. Transformation fails most often at the adoption stage, where employees are expected to change how they work but do not receive the ongoing support needed to sustain the change. In-workflow agents provide that support continuously, not as a one-time training event.
Closing the Gap, Not Just Talking About It
The training-to-performance gap is one of the oldest challenges in enterprise learning. It persists because the underlying architecture of most learning programmes creates distance between where development happens and where work happens.
In-workflow learning agents close that distance. They make development a continuous part of the work itself, rather than a separate activity that competes with it.
If your organisation is looking at how to embed learning agents into the workflows your teams already use, we can help you design and implement the approach.
Talk to our team at kydongrp.com/contact
Sources: Josh Bersin. "Learning in the Flow of Work." https://joshbersin.com/2018/06/a-new-paradigm-for-corporate-training-learning-in-the-flow-of-work/ Association for Talent Development. "2025 State of the Industry." https://www.td.org/research-reports Deloitte. "2025 Global Human Capital Trends." https://www2.deloitte.com/us/en/insights/focus/human-capital-trends.html