Artificial intelligence is a valuable business tool. Platforms such as ChatGPT and Claude can help generate ideas, review content, explain technical concepts, and offer another perspective.
I use AI regularly, and I genuinely appreciate what it can do. It helps me outline ideas, fine-tune copy when I cannot find the right words, develop creative concepts and imagery, and produce content in significantly less time. I love AI and welcome its continued evolution.
However, AI does not have human discernment.
It does not know the complete history of your brand, website, marketing strategy, or development process. It can analyze the information it is given, but it does not automatically understand the reasoning, experience, conversations, approvals, or technical considerations behind the work.
Depending on how a question is presented, AI may also reinforce what the user appears to want to hear rather than challenge the premise of the question.
That distinction matters.
A Real Example of AI Getting It Wrong
Recently, one of my clients ran completed artwork for a rack card and retractable floor banner through AI. The platform returned a long list of suggested changes. In my professional opinion, most—although not all—of those recommendations were unnecessary or lacked a legitimate marketing basis.
I still reviewed the feedback carefully. Full disclosure: I agreed that some of the copy could be reduced. As an advertising and marketing copywriter, I tend to begin with more content and edit it down during the design process. Guilty as charged. I love to write.
After the artwork was revised, I ran the updated pieces through AI again. It produced another list of entirely new recommendations. Many would have added more content, revisions, time, and cost without improving either piece.
I then told the AI that I disagreed. I did not provide new audience research, business information, or a detailed marketing explanation.
The platform immediately reversed its position without prompting. It acknowledged that its earlier recommendations had been excessive and concluded that the materials were fine as they were.
Think about that.
The files had not changed. No new information had been introduced. The only change was that I disagreed based on my professional experience.
The AI then adjusted its answer.
AI Can Sound Confident and Still Be Wrong
AI platforms often present recommendations in polished, confident language. That confidence can make an answer appear factual, researched, or authoritative when it may simply be a generalized suggestion.
AI is trained on enormous amounts of online content, which includes both accurate and inaccurate information. The average user may not always be able to recognize the difference, especially for technical issues.
More recommendations also do not automatically mean better marketing.
AI can review a file and generate observations. That is different from understanding the full strategy behind it.
AI Does Not Know Why a Marketing Decision Was Made
The same issue applies to website development and coding.
To date, we have run more than 80 websites through AI audits to evaluate what these tools could identify and how they can help us. AI provided a fast way to review certain front-end basics, which is super helpful, but it also made errors and incorrect assumptions.
Without professional experience, AI would not be able to recognize and correct those mistakes. Following every recommendation without review could have created new problems instead of solving existing ones.
Case in point, we recently had a client ask us to substantially reduce the copy on an e-commerce page based on AI feedback. However, that content had been intentionally developed as part of the page’s search engine optimization strategy.
Reducing the content without understanding its purpose could have weakened the page rather than improved it. Additional time was then required to revisit and re-review the strategy we had already discussed and confirmed, delaying completion.
Running the work through AI created unnecessary duplication and additional cost to the client for the time spent reviewing and responding.
Use AI to Ask Questions—not to Check Up on Your Experts
There is absolutely nothing wrong with using AI to become more informed or raise questions about your marketing.
There is, however, a major difference between saying:
“I ran this in AI and it raised this issue. What do you think?”
and:
“Please take this list from Chat GPT and make the changes.”
The first approach opens a productive professional discussion. The second begins with an assumption that may be incorrect and can be detrimental to the professional relationship you have established with your marketing partner.
Marketing Decisions Do Not Exist in Isolation
Copy affects layout. Layout affects readability. Readability affects response. Website features affect speed and performance. Code affects integrations. Tracking systems affect privacy and functionality. Development timing affects testing.
An AI recommendation may sound reasonable when viewed by itself while being completely wrong for the larger goal.
That is where professional experience matters.
Qualified professionals know when a recommendation will strengthen the work, when it will simply create more revisions, cost you more money and when it could weaken the final result.
AI Is a Tool, Not a Person
AI can be an excellent assistant, but it also makes mistakes, lacks project context, and may change its conclusions based on how the user responds rather than because the facts have changed.
Use AI. Ask thoughtful questions. Stay engaged in your marketing. But allow the professionals you hired to explain why an AI recommendation may be inaccurate, irrelevant, or potentially harmful.
Continually second-guessing professional decisions through AI can create confusion, duplicate work, increase costs, and damage an otherwise productive relationship.
The best marketing combines useful technology, qualified human judgment, and genuine client collaboration.
Until next time, keep thinking strategically.
Nella DeCesare
Managing Director & CEO
