ATD Blog
AI Will Not Replace L&D. But It Will Replace Order-Taking L&D
Content
AI is pushing TD leaders to re-evaluate every function against a harder standard.
AI is pushing TD leaders to re-evaluate every function against a harder standard.
Tue May 05 2026
Content
Generative AI is not the greatest threat to talent development. A narrow operating model is.
Generative AI is not the greatest threat to talent development. A narrow operating model is.
Content
As content production becomes faster and cheaper, L&D teams that still define their value as taking requests, building content, and tracking completions are becoming easier to question. The teams that will grow in relevance are the ones that are trusted partners to drive performance , thereby interpreting how work is changing, redesigning capability pathways, and building the human judgment that becomes more valuable as AI absorbs more routine tasks.
As content production becomes faster and cheaper, L&D teams that still define their value as taking requests, building content, and tracking completions are becoming easier to question. The teams that will grow in relevance are the ones that are trusted partners to drive performance, thereby interpreting how work is changing, redesigning capability pathways, and building the human judgment that becomes more valuable as AI absorbs more routine tasks.
Content
“If our value is just taking requests and producing learning assets faster, we are on borrowed time,” an L&D executive recently told me.
“If our value is just taking requests and producing learning assets faster, we are on borrowed time,” an L&D executive recently told me.
Content
For years, talent development has lived with a contradiction. Leaders routinely say people are their greatest asset, yet many learning teams are still funded and evaluated like internal production shops. Build the module. Launch the workshop. Create the job aid. Track completions. Move on.
For years, talent development has lived with a contradiction. Leaders routinely say people are their greatest asset, yet many learning teams are still funded and evaluated like internal production shops. Build the module. Launch the workshop. Create the job aid. Track completions. Move on.
Content
Generative AI has exposed the weakness of that model faster than many expected. A line of business leaders can now look at tools that generate slide content, scripts, videos, voiceover, quizzes, summaries, and first-draft learning assets in minutes and ask a hard question: If technology can produce so much of what L&D has historically been asked to create, what is the function for now ?
Generative AI has exposed the weakness of that model faster than many expected. A line of business leaders can now look at tools that generate slide content, scripts, videos, voiceover, quizzes, summaries, and first-draft learning assets in minutes and ask a hard question: If technology can produce so much of what L&D has historically been asked to create, what is the function for now?
Content
That question can make people in our field uncomfortable. It can also be clarifying. AI is not making talent development irrelevant. But it is making order-taking talent development irrelevant.
That question can make people in our field uncomfortable. It can also be clarifying. AI is not making talent development irrelevant. But it is making order-taking talent development irrelevant.
Content
The issue is not whether learning still matters. It does. The issue is whether the old value proposition still holds. And in many organizations, it does not.
The issue is not whether learning still matters. It does. The issue is whether the old value proposition still holds. And in many organizations, it does not.
The real threat is not AI. It is a weak operating model.
Content
Too much of the AI discussion in L&D is still focused on the visible layer of change: prompt writing, faster content production, AI-generated videos, and how to make the same deliverables more efficiently. Some of that is useful. None of it gets to the center of the issue.
Too much of the AI discussion in L&D is still focused on the visible layer of change: prompt writing, faster content production, AI-generated videos, and how to make the same deliverables more efficiently. Some of that is useful. None of it gets to the center of the issue.
Content
The deeper issue is that AI is pushing leaders to re-evaluate every function against a harder standard. Does this team mainly produce outputs, or does it improve business performance?
The deeper issue is that AI is pushing leaders to re-evaluate every function against a harder standard. Does this team mainly produce outputs, or does it improve business performance?
Content
That is an uncomfortable standard for L&D because many teams have spent years being rewarded for responsiveness and volume. More courses. More assets. More requests fulfilled. More completions. Those measures were never strong proof of strategic value. Now that training production is becoming less scarce, they are becoming even less persuasive.
That is an uncomfortable standard for L&D because many teams have spent years being rewarded for responsiveness and volume. More courses. More assets. More requests fulfilled. More completions. Those measures were never strong proof of strategic value. Now that training production is becoming less scarce, they are becoming even less persuasive.
Content
McKinsey estimates that currently demonstrated technologies could, in theory, automate activities accounting for about 57 percent of US work hours today. At the same time, it estimates that more than 70 percent of today’s skills can still be applied in both automatable and non-automatable work. In other words, the work is not disappearing neatly. It is being reassembled, and people will still be central to making it work well.
McKinsey estimates that currently demonstrated technologies could, in theory, automate activities accounting for about 57 percent of US work hours today. At the same time, it estimates that more than 70 percent of today’s skills can still be applied in both automatable and non-automatable work. In other words, the work is not disappearing neatly. It is being reassembled, and people will still be central to making it work well.
Content
That is exactly why the old L&D model is under pressure. If work is being recombined, then the real value is no longer just producing learning materials. The real value is helping the organization understand what people now need to do differently and how they will build that capability.
That is exactly why the old L&D model is under pressure. If work is being recombined, then the real value is no longer just producing learning materials. The real value is helping the organization understand what people now need to do differently and how they will build that capability.
Most teams are responding to visible AI activity, not the quieter redesign of work.
Content
Many organizations are reacting to AI by focusing on what is easiest to see: vendor demos, pilots, policy documents, prompt training, and a growing list of tools. That activity can look like transformation because it is visible. But visible activity is not the same as structural change.
Many organizations are reacting to AI by focusing on what is easiest to see: vendor demos, pilots, policy documents, prompt training, and a growing list of tools. That activity can look like transformation because it is visible. But visible activity is not the same as structural change.
Content
What matters more is the quieter redesign happening underneath. Tasks are being redistributed. Some routine work is being delegated to AI. Some roles are becoming more judgment-heavy. Managers are spending less time reviewing raw output and more time validating, coaching, and orchestrating work across people and systems. Governance, verification, and oversight are becoming increasingly important.
What matters more is the quieter redesign happening underneath. Tasks are being redistributed. Some routine work is being delegated to AI. Some roles are becoming more judgment-heavy. Managers are spending less time reviewing raw output and more time validating, coaching, and orchestrating work across people and systems. Governance, verification, and oversight are becoming increasingly important.
Content
Evidence from a recent, large-scale, task‑level analysis is useful because it captures the less visible layer of work change. Drawing on 531,504 task-level signals, it found that 37 percent of work was delegable, 63.2 percent remained fundamentally human-led, and 14.3 percent of tasks did not map cleanly to legacy job families at all. That last number matters. It suggests that hybrid work patterns are already forming outside old role definitions.
Evidence from a recent, large-scale, task‑level analysis is useful because it captures the less visible layer of work change. Drawing on 531,504 task-level signals, it found that 37 percent of work was delegable, 63.2 percent remained fundamentally human-led, and 14.3 percent of tasks did not map cleanly to legacy job families at all. That last number matters. It suggests that hybrid work patterns are already forming outside old role definitions.
Content
For L&D, this is a warning. If we respond only to the visible side of AI, we will focus on training tools for our workforce. If we pay attention to the quieter redesign of work, we will ask better questions. Where is human judgment becoming more valuable? Which roles are changing before job descriptions catch up? Where are managers being asked to make better decisions, not just faster ones? Where does the business need workflow redesign rather than one more course? That’s where our value as a profession emerges.
For L&D, this is a warning. If we respond only to the visible side of AI, we will focus on training tools for our workforce. If we pay attention to the quieter redesign of work, we will ask better questions. Where is human judgment becoming more valuable? Which roles are changing before job descriptions catch up? Where are managers being asked to make better decisions, not just faster ones? Where does the business need workflow redesign rather than one more course? That’s where our value as a profession emerges.
The labor signal L&D should not ignore.
Content
There is still too much binary thinking about AI and jobs. Either people predict a near-term collapse in employment, or they dismiss the issue because mass layoffs have not appeared everywhere at once. The evidence points to something more nuanced.
There is still too much binary thinking about AI and jobs. Either people predict a near-term collapse in employment, or they dismiss the issue because mass layoffs have not appeared everywhere at once. The evidence points to something more nuanced.
Content
Anthropic’s labor-market research found no clear rise in unemployment for workers in the most AI-exposed jobs since late 2022. For many leaders, that may sound reassuring.
Anthropic’s labor-market research found no clear rise in unemployment for workers in the most AI-exposed jobs since late 2022. For many leaders, that may sound reassuring.
Content
But that is not the whole story. The more revealing signal may be at the front end of the labor market. Anthropic found that for workers ages 22 to 25, entry into the most exposed occupations fell by about half a percentage point per month, which the report estimates as a 14 percent drop in job-finding rates compared with 2022 levels.
But that is not the whole story. The more revealing signal may be at the front end of the labor market. Anthropic found that for workers ages 22 to 25, entry into the most exposed occupations fell by about half a percentage point per month, which the report estimates as a 14 percent drop in job-finding rates compared with 2022 levels.
Content
That finding should matter to talent development far more than it currently does. If entry-level pathways narrow, organizations can lose one of the main ways people have historically developed judgment: experience. Many future managers built capability through frontline and early-career work and gradual exposure to complexity. If AI removes some of that proving-ground work, the leadership pipeline does not stay intact on its own.
That finding should matter to talent development far more than it currently does. If entry-level pathways narrow, organizations can lose one of the main ways people have historically developed judgment: experience. Many future managers built capability through frontline and early-career work and gradual exposure to complexity. If AI removes some of that proving-ground work, the leadership pipeline does not stay intact on its own.
Content
This is where L&D has an opportunity to become more important, not less. As informal development through experience becomes less reliable, organizations will need more intentional, better-designed practice , coaching, and capability design. They will need stronger pathways for judgment, decision making, communication, and applied problem solving. They will need learning leaders who can see the developmental implications of work redesign before the talent pipeline weakens in obvious ways.
This is where L&D has an opportunity to become more important, not less. As informal development through experience becomes less reliable, organizations will need more intentional, better-designed practice, coaching, and capability design. They will need stronger pathways for judgment, decision making, communication, and applied problem solving. They will need learning leaders who can see the developmental implications of work redesign before the talent pipeline weakens in obvious ways.
The old value proposition of L&D is weakening.
Content
This is the point many in our field still resist. We are not simply seeing new tools . We are seeing the erosion of an older claim to value.
This is the point many in our field still resist. We are not simply seeing new tools. We are seeing the erosion of an older claim to value.
Content
For a long time, L&D could justify itself in part by being one of the few internal teams equipped to turn subject matter into organized learning experiences. That still matters. Strong design still matters. Context still matters. Good judgment still matters. But generative AI has reduced the scarcity of training production.
For a long time, L&D could justify itself in part by being one of the few internal teams equipped to turn subject matter into organized learning experiences. That still matters. Strong design still matters. Context still matters. Good judgment still matters. But generative AI has reduced the scarcity of training production.
Content
If a function still defines itself mainly as the group that converts information into training assets, line-of-business leaders will continue to ask whether a smaller team using better tools can do the same work faster. In some organizations, they will decide the answer is yes.
If a function still defines itself mainly as the group that converts information into training assets, line-of-business leaders will continue to ask whether a smaller team using better tools can do the same work faster. In some organizations, they will decide the answer is yes.
Content
That’s why this moment calls for a more disciplined definition of value. L&D earns relevance when it helps the business anticipate capability gaps, strengthen managers through changing work, and build the skills that are becoming more valuable as routine tasks are absorbed, accelerated, or redistributed.
That’s why this moment calls for a more disciplined definition of value. L&D earns relevance when it helps the business anticipate capability gaps, strengthen managers through changing work, and build the skills that are becoming more valuable as routine tasks are absorbed, accelerated, or redistributed.
Content
McKinsey’s findings reinforce this point. Demand for AI fluency has jumped nearly sevenfold in two years, while people-focused skills such as coaching and negotiation show among the least exposure to automation. Routine skills such as invoicing are among the most. That is not a sentimental claim that “humans still matter.” It is a practical signal about where capability investment is likely to hold more value over time.
McKinsey’s findings reinforce this point. Demand for AI fluency has jumped nearly sevenfold in two years, while people-focused skills such as coaching and negotiation show among the least exposure to automation. Routine skills such as invoicing are among the most. That is not a sentimental claim that “humans still matter.” It is a practical signal about where capability investment is likely to hold more value over time.
What strategic L&D looks like now.
Content
The strongest learning teams are moving away from a production-centered identity and toward a performance-centered one.
The strongest learning teams are moving away from a production-centered identity and toward a performance-centered one.
Content
They are not waiting to be handed a training request. They are working earlier in the cycle, when work itself is being reconfigured . They are asking where AI is changing tasks, which capabilities are becoming more valuable, what managers now need from their teams, and where the organization will struggle if capability development remains tied to the old shape of work.
They are not waiting to be handed a training request. They are working earlier in the cycle, when work itself is being reconfigured. They are asking where AI is changing tasks, which capabilities are becoming more valuable, what managers now need from their teams, and where the organization will struggle if capability development remains tied to the old shape of work.
Content
They also understand that AI readiness is not just a learning team issue. It is a workforce issue. These teams recognize that the workforce is at risk of cognitive decline and are helping their organizations adapt for the future. McKinsey notes that organizations will need more than technical knowledge. They will need adaptability, critical thinking, collaboration, analytical judgment, and the ability to question outputs and recognize bias or error. That sits squarely in talent development’s lane.
They also understand that AI readiness is not just a learning team issue. It is a workforce issue. These teams recognize that the workforce is at risk of cognitive decline and are helping their organizations adapt for the future. McKinsey notes that organizations will need more than technical knowledge. They will need adaptability, critical thinking, collaboration, analytical judgment, and the ability to question outputs and recognize bias or error. That sits squarely in talent development’s lane.
Content
In practical terms, strategic L&D now has at least four jobs.
In practical terms, strategic L&D now has at least four jobs.
Content
Interpret business change as capability change.
Interpret business change as capability change.
Content
Help redesign learning and practice pathways where frontline work no longer provides enough foundational experience for development.
Help redesign learning and practice pathways where frontline work no longer provides enough foundational experience for development.
Content
Strengthen managers so they can coach, verify, and exercise judgment in AI-shaped workflows.
Strengthen managers so they can coach, verify, and exercise judgment in AI-shaped workflows.
Content
Show value in business terms, not just activity metrics.
Show value in business terms, not just activity metrics.
Content
That last point is especially important. Completion rates, attendance, and satisfaction scores can still be useful, but they are weak signals when budgets tighten. As training creation becomes easier to automate, learning teams will increasingly be judged by a tougher question: What changed because this team exists?
That last point is especially important. Completion rates, attendance, and satisfaction scores can still be useful, but they are weak signals when budgets tighten. As training creation becomes easier to automate, learning teams will increasingly be judged by a tougher question: What changed because this team exists?
Content
That is a fair question. The teams that can answer it in business language will have a stronger future than the teams that answer it with volume.
That is a fair question. The teams that can answer it in business language will have a stronger future than the teams that answer it with volume.
It’s a harder and better future for the profession.
Content
There is a temptation in our field to frame every major change as an opportunity without acknowledging the threat inside it. This moment is both.
There is a temptation in our field to frame every major change as an opportunity without acknowledging the threat inside it. This moment is both.
Content
Yes, AI gives talent development a chance to lead on workforce readiness, capability building, manager enablement, and more intentional skill development. It also threatens any version of the function that still defines itself too narrowly.
Yes, AI gives talent development a chance to lead on workforce readiness, capability building, manager enablement, and more intentional skill development. It also threatens any version of the function that still defines itself too narrowly.
Content
The old compact is weakening. “We create learning materials, therefore we are essential” is no longer a sufficient case. AI can now perform parts of that work at a speed leaders can plainly see. Meanwhile, the stronger research suggests that the real effects of AI may first show up through changes in skill demand, thinner entry pathways, and quieter redesign of work rather than through immediate, visible job loss.
The old compact is weakening. “We create learning materials, therefore we are essential” is no longer a sufficient case. AI can now perform parts of that work at a speed leaders can plainly see. Meanwhile, the stronger research suggests that the real effects of AI may first show up through changes in skill demand, thinner entry pathways, and quieter redesign of work rather than through immediate, visible job loss.
Content
That leaves L&D with a choice. We can use AI to become a faster “content factory,” or we can use this moment to become more strategically useful .
That leaves L&D with a choice. We can use AI to become a faster “content factory,” or we can use this moment to become more strategically useful.
Content
The second path is harder. It requires sharper business judgment, more confidence, and a willingness to move upstream from delivery to diagnosis. But it is also the path that gives the profession a more credible and durable future.
The second path is harder. It requires sharper business judgment, more confidence, and a willingness to move upstream from delivery to diagnosis. But it is also the path that gives the profession a more credible and durable future.
Content
AI will not replace L&D. But it will replace order-taking L&D . Our opportunity now is to leave that version behind and step more fully into the work that has always mattered most: helping our organizations perform better through people.
AI will not replace L&D. But it will replace order-taking L&D. Our opportunity now is to leave that version behind and step more fully into the work that has always mattered most: helping our organizations perform better through people.
Content
Enjoyed this article? Check out these AI resources:
Enjoyed this article? Check out these AI resources:
Content
ATD Research: How Instructional Designers Use AI
Content
Become a Strategic Learning Architect With AI
Content
Beyond the Prompt: Why Ohio State’s Master of Learning Technologies Teaches Analysis Over Automation
Beyond the Prompt: Why Ohio State’s Master of Learning Technologies Teaches Analysis Over Automation
Content