The Expert Eye: Why Experience Still Matters in the Age of AI
My high school art teacher had a gift I didn't understand until years later. Every time I thought a piece was finished, he'd lean over my shoulder and say, "Almost. Just needs a little more." I'd stare at my work, baffled. What did he see that I didn't?
He had the eye of an expert and could see what was lacking, missing or not quite right in a piece of art. I didn't have that “eye”…at least not yet.
That gap between good and great is invisible to a novice. And if you have been working in your craft for a while — whether’s is communications, design, project management or any other area of expertise — that expertise of “knowing” the difference between done and exceptional only becomes visible through years of looking, making, failing, and refining.
And it turns out, that expertise gap is also where AI runs into its limits.
AI Didn't Take Our Jobs. But It Did Start a Very Interesting Conversation.
I was on a call last week with a good friend. She’s a scientific illustrator and an old colleague of mine. We were catching up, but the conversation kept going back to the same question: Is AI going to take my job?
We're in different fields. She illustrates. I write and do marketing and communications strategy. But regardless, the answer is the same.
AI will not take your job if you are an expert in your field and know what you're doing.
Here's what I mean.
I recently tried to get Claude (sorry, Claude) to design a LinkedIn ad for a campaign I was running. I gave it headshots. I gave it the color palette. I spelled out exactly what I wanted. The result was, to put it charitably, not good. I ended up scrapping it and doing it myself in Canva in half the time. Obviously, I’m not anti-AI, but because I could see that it was not what my client would want, and I had the skills to fix it without burning another hour re-prompting a machine.
AI is wonderful, but the tools can’t replace creative and strategic judgment, which takes years to build.
The Difference Between Using AI and Knowing What to Do With It
My friend described her current process when using AI tools. She says she sketches something by hand, uses AI to help refine the style, then reworks it herself in Illustrator. While she uses AI to assit, it’s never Claude or Gemini or any other design bot leading the way. Her vision leads.
As an artist ,she knows what she wants. “I can explain what I want more carefully to the LLM, tease the result a bit and make it more original. This is something that makes it easier for me than someone without the years of experience or background," she said.
Someone without design experience can use the same tools and get something that looks... fine. It’s totally passable, but also forgettable. Because they don't know what's missing. They can't feel when something is off. They don't see that the composition pulls the eye in the wrong direction, or that the color palette is screaming "generic template" or that the video looks just like everyone else’s video on social media (yes, people are already getting bored of the Capcut templates).
When Illustrator launched in 1987, it didn't make everyone a designer. Dreamweaver (which is most famous for popularizing the WYSIWYG , the “what you see is what you get” web design tool) didn't make everyone a web developer. AI won't make everyone a writer or an artist, but it will make the gap between novice output and expert output less immediately obvious.
Until someone with the “expert eye” takes a look.
The Skill AI Will Never Be Able to Prompt For
And there's something else AI cannot simulate and that is working with humans. If you’ve worked in an agency, an art department or even a corporate communications team, you know your client or stakeholders don’t always know exactly what they want. They might have a rough idea or they might have a goal in mind, but the details aren’t there or they aren’t quite sure exactly what they need yet.
So much of what I do and what my friend does is essentially mind reading. When we sit with a client, we get to understand their personality and their expectations in a way that can’t always be translated in an email. People express themselves through both body language and words – and sometimes it’s what they don’t say that makes a difference.
You have to ask the right questions, try things, and show direction until something clicks. That's not something that can’t be prompted. That's years of working with various clients and gaining hard-won instincts.
The other things is when the AI runs out of tokens and gives up mid-project, which, trust me, happens, the person with experience can keep going. They can take what they’ve got and finish it themselves. Someone who is relying entirely on the tool? They're stuck.
Experience Is the Thing AI Can't Generate
AI is an accelerant. It makes me faster. It enables me to generate first drafts and variations. It helps me brainstorm and test out ideas. It can analyze all the data I drop in quickly. From there, I can use my expertise to draw deeper insights and really be strategic with my client’s plans.
Then there is agentic AI that helps me monitor media mentions, flag possible events and speaking engagments for my clients and track their competitors.
But even though AI is wonderful, it still needs someone at the wheel who knows the difference between what almost looks like and what it actually takes to get to great. It needs someone with expertise to challenge it…someone who can fine-tune results and make them even better and more relevant.
So are all you experts out there going to be replaced by AI? Nope.
Your judgement is needed now more than ever to help clients stand out.
In fact, rather than take your job, I foresee AI becoming an essential tool in your toolkit — one that allows you to do even more fantastic work.