Imagine walking into your doctor’s office and saying:
“I asked AI about my symptoms. It says I have a liver problem. I’d like medication.”
The doctor smiles and replies:
“Of course. Which medication were you thinking?”
No examination.
No blood work.
No questions.
No attempt to determine whether the liver is actually the problem.
We would call that malpractice.
Yet in Learning & Development, we do the equivalent every single day.
A manager says:
“We need training.”
And our response is often:
“Sure. When do you need it?”
No diagnosis.
No investigation.
No evidence that training is even the right intervention.
Somehow, we’ve accepted that as normal.
AI Didn’t Create This Problem. It Exposed It.
Artificial intelligence is changing how people seek advice.
Patients research symptoms before visiting their doctor.
Investors ask AI about retirement strategies before meeting a financial advisor.
Homebuyers ask AI which neighbourhoods are “best” before calling a Realtor.
Employees ask AI how to improve leadership, communication, or customer service before speaking to Learning & Development.
None of that is inherently bad.
In fact, it’s often helpful.
People are arriving with ideas, questions, and possibilities.
The problem begins when those ideas become prescriptions.
AI can generate remarkably plausible solutions.
What it cannot do is determine whether it is solving the right problem.
That’s where professionals earn their value.
The Difference Between Information and Diagnosis
Doctors don’t object when patients do research.
Good doctors welcome it.
The patient brings observations, concerns, family history, internet research, and perhaps an AI-generated explanation.
The doctor treats all of that as input—not as proof.
Then the diagnostic process begins.
Tests are ordered.
Alternative explanations are considered.
Evidence is gathered.
Only then does treatment begin.
Notice the order.
Diagnosis comes before prescription.
Every time.
L&D Often Works Backwards
Consider how many conversations begin inside organizations.
“We need customer service training.”
“We need leadership training.”
“We need a communication workshop.”
“We need an e-learning module.”
Sometimes those requests are now even more polished.
“I asked ChatGPT. It recommended a blended learning pathway with simulations, microlearning, manager coaching, and reinforcement.”
It sounds sophisticated.
It sounds evidence-based.
But it’s still a prescription.
No one has established whether the problem is actually capability.
Perhaps expectations are unclear.
Perhaps the process is broken.
Perhaps managers are reinforcing the wrong behaviours.
Perhaps the technology creates unnecessary friction.
Perhaps incentives reward the wrong outcomes.
Perhaps people already know what to do but cannot do it because the environment makes good performance impossible.
None of those problems are solved by better training.
Yet training is often the first—and sometimes only—solution considered.
Why Does This Keep Happening?
I don’t believe L&D lacks intelligence.
I don’t believe practitioners lack good intentions.
I think we’ve inherited a role that encourages production more than diagnosis.
Organizations ask us to build.
Vendors sell solutions.
Success is measured by courses delivered, completions achieved, satisfaction scores, and utilization.
Very little asks whether the intervention should have existed in the first place.
Imagine if medicine measured physicians by:
- Number of prescriptions written
- Number of surgeries performed
- Number of diagnostic tests ordered
Rather than patient outcomes.
We would expect overtreatment.
L&D has unintentionally created a similar incentive system.
The safest response is often to build something.
Not to question whether anything should be built at all.
AI Makes This More Important, Not Less
Some people worry AI will replace advisory professions.
I think it changes them.
The role is no longer to be the person with the information.
Information is becoming abundant.
Judgment is becoming scarce.
The value of a doctor isn’t remembering every possible disease.
It’s knowing which diagnosis fits this patient.
The value of a financial advisor isn’t knowing every investment product.
It’s understanding which strategy fits this family.
The value of a lawyer isn’t having access to legal information.
It’s knowing how the law applies in this situation.
And the value of Learning & Development isn’t producing learning solutions.
It’s determining whether learning is required at all.
That’s a very different profession.
Don’t Stop Stakeholders From Using AI
Quite the opposite.
Encourage it.
Ask stakeholders to bring everything AI suggested.
Bring the articles.
Bring the frameworks.
Bring the recommendations.
Bring the brainstorming.
Then treat those ideas exactly as a doctor treats a patient’s internet research:
As useful input.
Not as the diagnosis.
The conversation changes from:
“What do you want us to build?”
to
“What evidence tells us this is the problem?”
Only after answering that question should we ask:
“Now, what is the most appropriate intervention?”
Sometimes it will be training.
Sometimes coaching.
Sometimes process redesign.
Sometimes performance support.
Sometimes better tools.
Sometimes leadership.
Sometimes no intervention at all.
If We Want to Be Seen as Professionals…
Many in L&D want a stronger strategic voice.
A seat at the executive table.
Greater credibility.
More influence.
Those aspirations are entirely reasonable.
But they require something more than better learning design.
They require professional discipline.
Doctors diagnose before they prescribe.
Lawyers investigate before they advise.
Engineers analyze before they design.
Auditors gather evidence before they conclude.
Management consultants study the system before recommending change.
Perhaps Learning & Development should hold itself to the same standard.
Because organizations don’t need us simply to build better training.
They need us to make better decisions about whether training belongs in the solution at all.
AI hasn’t made diagnosis less important.
It has made it indispensable.
The future of Learning & Development will not belong to the teams that generate the fastest solutions.
It will belong to the professionals who refuse to prescribe until they understand the problem.
