AI can now build a decent presentation in less time than it takes to refill your coffee.
Give it a report, meeting notes, and an old deck you like. A few minutes later, you may have a polished set of slides with charts, colors, layouts, and formatting that look surprisingly good.
That is impressive. It is also a little distracting.
I recently watched an episode of Next Slide, Please about new presentation features in Google’s Gemini and Anthropic’s Claude. The hosts showed these tools using existing decks as visual references. The AI did not just grab similar colors. It copied small design choices such as spacing, margins, type styles, and layout patterns.
The obvious story is that AI is getting better at making slides.
The more important story is that making slides is becoming less valuable than deciding what belongs on them.
Production is getting cheap
For years, many knowledge workers have been judged by what they produce.
How many courses did you build? How many slides did you create? How many reports, job aids, videos, or campaigns did your team ship?
Those numbers are easy to see, so they are easy to reward. But they have never been the same as value.
Now AI is exposing that gap.
When almost anyone can create ten versions in minutes, the important question is no longer, “Can you make this?”
It is, “Can you tell me which version is worth using, and why?”
That is judgment.
The real work was hiding inside the work
Most people think building a presentation is about writing slides and arranging objects. Some of it is. But the harder work happens before the final layout.
You sort through too much information. You decide what the audience needs and what it does not. You test the logic, cut weak points, question assumptions, and choose what deserves attention.
The deck is the visible artifact. Judgment is the work underneath it.
The same is true when you design training, write an article, plan a campaign, or recommend a solution. The finished product gets noticed. The choices behind it create the value.
AI can remove a lot of production work, and I am happy to let it. I have no deep emotional attachment to resizing text boxes.
But if we hand off the choices too, we are not saving time. We are giving away the part of the job that matters most.
Polish can hide weak thinking
This is where things get tricky.
In the past, messy thinking often produced messy slides. The story wandered. The design felt crowded. The gaps were easier to see.
AI can now wrap weak thinking in a clean layout.
A shaky claim can sit beside a polished chart. A shallow analysis can look finished. A recommendation built on thin evidence can arrive with confident headings, tasteful colors, and very tidy margins.
The better AI gets at presentation, the easier it becomes to confuse polish with quality.
They are not the same.
A useful test is to imagine the meeting after the deck is done. Someone asks where a number came from. A leader challenges the recommendation. A subject matter expert points out the exception that changes everything.
The slides cannot defend the thinking. You have to.
When content becomes abundant, judgment becomes scarce
AI is not only changing presentation design. It is changing learning and development, marketing, consulting, research, and nearly every kind of knowledge work.
Content is becoming abundant. Judgment is not.
Judgment means weighing the evidence, reading the context, recognizing the tradeoffs, and making a choice you can explain. It requires a different set of questions:
- Is this the right problem to solve?
- What evidence supports this conclusion?
- What assumptions are we making?
- What might the AI have missed or flattened?
- Which option best fits this audience and situation?
- What should happen next?
AI can generate options. It can summarize the inputs and suggest a path. It may even spot something you missed.
But it does not know your organization, audience, constraints, and stakes the way you do. More importantly, it cannot own the consequences of the choice.
That part is still ours.
This is an opportunity for L&D
For years, learning professionals have talked about becoming less like order takers and more like performance consultants. AI may force that shift faster than any new model or framework ever did.
Imagine a leader asks for training because a team keeps missing a step in an important process. AI can draft the course, create a quiz, write a job aid, and produce the launch email before lunch.
Useful? Absolutely.
But none of those assets matter if the real problem is a confusing system, a broken process, competing priorities, or a step that should not exist in the first place.
AI can help us investigate the problem. It can suggest questions, compare possible causes, and organize the evidence. It should not get the final vote.
That is where learning professionals can bring more value.
Our best contribution is not the number of courses or slides we produce. It is diagnosing the problem before prescribing training. It is asking better questions, making sense of conflicting signals, earning a stakeholder’s trust, and connecting learning to real performance.
When production gets cheaper, those skills become easier to see.
Use AI for leverage, not permission
I am not arguing that we should slow AI down or cling to every manual task. Quite the opposite.
Let AI create the first draft. Let it clean up the deck, summarize the research, suggest a structure, and give you options. Then use the saved time to do the work that deserves more attention.
Before you put your name on the result, ask:
- Can I explain the reasoning in plain language?
- Can I trace the important claims back to reliable evidence?
- Do I understand the tradeoffs?
- What would make me change my recommendation?
- Will this help someone make a better decision or improve performance?
If you cannot answer those questions, the work is not finished, no matter how polished it looks.
The bottom line
AI is lowering the cost of creating content. That is useful, exciting, and probably overdue.
But lower production costs do not remove the need for expertise. They reveal where expertise was hiding all along.
It was never just in making the thing.
It was in knowing what to make, what to leave out, what to question, and what to do next.
When AI can make almost anything, judgment becomes the job.