Bar Maimon, AI Growth Hacker, Founder, and CEO at IceWeb.
According to Ahrefs, 74% of new webpages contain AI-generated copy, with over one-third of this content containing 41% or more AI content. To boost their SEO results, many brands are won over by how easy it is to write and publish AI content.
Yet both readers and search engines know something is off. While the words are technically correct, the content feels flat and unconvincing. It reads exactly like your competitors’. And it fails to generate traffic or bring in leads. Worse, search engines may demote your site if it lacks original insights. In Grokipedia’s example, the site’s traffic plummeted between January and February 2026.
As with many technologies, the problem doesn’t lie solely with AI itself. It also lies with people.
The Mistake Everyone Makes
The first question a good content writer asks is “Why am I writing this content?” followed by “Who am I writing for?”
A generic text written by an LLM may struggle to answer such basic questions. Based on the Ahrefs statistics, many people open an AI search engine, type “Write me a blog post about X” and hit publish after a light edit. The result is repetitive, bland content that covers the topic without saying anything—and that may be wildly inaccurate.
This approach treats AI as a ghostwriter, which is the wrong job description entirely. Something I’ve learned as a founder in digital marketing is that a good content writer should align their content with the brand’s voice, audience and experience. The content’s purpose is to deliver information, convince readers of a brand’s authority, bring in leads and raise its website’s SEO ranking.
An LLM struggles to do all but one of these things. Even if it produces a reasonable first draft, polishing that into original content still takes time and experience. AI does, however, possess an extraordinary ability to process, organize and structure information. That’s a completely different skill set, and one that’s genuinely useful once you stop asking it to do something it was never built to do.
What AI Is Actually Good At
Think of an LLM as the world’s best-read research assistant that can scan and combine tons of data to deliver information within seconds. It also just so happens that it has no personality whatsoever. The voice and nuances of content must come from the prompt and the writer.
Ask it to map out every angle a topic could be approached from, and it can give you a dozen ideas you may not have considered because it crawled through thousands of websites and pages of online content, assimilated all this information and condensed it into a handful of useful bullet points. Instruct it to build a content structure, spot gaps in your argument or suggest what questions your target audience is asking, and it can deliver real value. Ask it to double-check your facts, and it should point out mistakes in its own answers. A well-designed LLM built on solid data can do all of this because it can sift through data at speeds that no human brain could ever achieve.
An AI tool can generate a first-pass outline, deliver section headers or flag logical and factual errors before you’ve written a single paragraph. It’s a helpful thinking partner that can help accelerate the structural work.
What Only You Can Do
Since they lack real-world experience, LLMs can’t replicate a content writer’s experience or the nuances that make up a business or a brand. Writers know how to frame a problem to highlight a specific angle and guide the audience toward a specific goal. Voice and judgment matter immensely in content writing because a writer’s experience tells them which points land hardest with an audience and how to guide visitors through the buying journey with the right arguments.
The Workflow That Actually Works
In my experience, the best content happens through a division of labor that looks like this:
1. Use the LLM to test your topic.
Before writing anything, ask the LLM what the standard takes on this subject are, so you can ensure your content is interesting, novel and original. You don’t want to compete with thousands of pages of similar content, or your content will rank low on SEO. Ask what long-tail questions someone new to the topic would have and what a skeptic might push back on. Use those answers to define your angle.
2. Ask for an outline.
Next, ask for an outline that contains the main ideas, sections and sub-sections based on the information you’ve decided to move forward with. Review it, move things around, cut what doesn’t serve your purpose and add sections as necessary. At this stage, the AI has done the scaffolding work and your judgment has shaped it into something with a point of view.
3. Write your version of the content.
It’s now time to shine through your writing with your own words and point of view. Using the LLM structure as your guide, fill in content with a voice that matches the organization and its goals. Never forget the key question, “Why am I writing this content?” Adjust your voice, writing and examples to the article’s overall purpose.
Once you have your first draft, loop it back into the LLM to check for weak arguments or factual errors. Use the feedback to polish your article, perhaps even creating your own acronyms where appropriate.
Final Thoughts
Both readers and search engines are getting better at detecting content written to fill space rather than to tell a story or help them. Many know when a text is generic and superficial—another drop in an ocean of AI slop. Those readers won’t read it, nor will search engines index it.
I believe the path to staying ahead on content over the next few years involves using AI to kick-start content that you then tailor to your audience, written with purpose and backed by credibility and self-confidence. AI can make you a faster, sharper thinker, providing you with novel ideas. But the writing? That’s still all yours.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
