I Built My Ad Strategy Using AI — Here's What It Got Wrong
I Built My Entire Blog's Ad Strategy Using AI — Here's What It Got Wrong
Fikrago exists because of AI. That's not an exaggeration or a marketing line — I have no coding background, no budget, and no team. Every tool, every article, every piece of the site's technical foundation got built with AI doing the heavy lifting alongside me. So when it came time to plan out a monetization strategy — how many articles to publish, how to structure them, what would actually get me approved by an ad network — leaning on AI for that too felt like the obvious next step. It had gotten me this far. Why wouldn't it get me through this part too?
It was low value, and now I need to do articles with real value — that's the blunt way I'd sum up what actually happened. And figuring out why the strategy went sideways taught me something important about where AI genuinely helps a solo builder, and where it quietly leads you astray if you're not watching closely enough.
What the Strategy Actually Looked Like
The early version of my content plan, shaped heavily by AI suggestions, prioritized volume and structure over almost everything else. Publish a lot. Hit word counts. Cover a wide range of topics inside the general "make money online" space. Follow SEO best practices — title tags, meta descriptions, keyword placement — and the traffic and approval would follow. On paper, that's a completely reasonable strategy. It's the kind of advice you'd find in a dozen legitimate guides about growing a blog.
The problem wasn't that the advice was wrong in a vacuum. It's that AI, working from general patterns about what "good content" looks like, optimized for the version of quality that's measurable and structural — length, formatting, keyword density — without any real ability to judge whether the content underneath all that structure actually had a genuine, personal, hard-to-replicate point of view behind it. It could tell me an article should be 1,000 words. It couldn't tell me whether those 1,000 words sounded like they came from an actual person who had lived through something, or from a competent summary of information that existed elsewhere already.
The Mistake That Mattered Most
The clearest mistake, looking back, was treating "publish more, cover more ground" as the core growth lever. AI is very good at generating plausible-sounding content across a wide range of topics quickly, and that capability quietly pulled my strategy toward breadth instead of depth — more articles touching more angles of the make-money-online space, rather than fewer articles going deeper into the specific, personal, case-study style story that only I could actually tell.
That approach produced a site with real volume but uneven substance. Some articles genuinely had a personal hook and a real story behind them. Others were closer to competent, well-structured summaries — technically correct, technically optimized, but replaceable by any other blog covering the same topic in the same generic way. And that second category is exactly what "low value content" describes, once you strip away the discomfort of the label and look at what it's actually pointing at. Not bad writing. Generic writing. Content that could have come from anywhere, attached to a site that's supposed to be about one specific person's real, unfiltered build-in-public story.
The Defense I Considered
There's a version of this where I blame the tool entirely — AI led me down the wrong path, and the fix is simply to use it less. I don't think that's accurate, and I don't think it's fair either. AI didn't write anything without me approving it, prompting it, and publishing it. The strategic mistake wasn't trusting AI to help — it was trusting AI's version of "quality" as if it were the same thing as my version of "value," without checking closely enough whether those two things actually lined up.
AI's version of quality, left unchecked, defaults to the measurable stuff: structure, length, keyword coverage, formatting compliance. My version of value — the thing that actually makes Fikrago worth reading over any other AI-and-business blog — is the personal story underneath the structure. The rejection, the empty platform, the late nights, the real reactions to real setbacks. AI can help me organize and sharpen that story once I bring it. It can't manufacture the story itself, and my early strategy asked it to try anyway, by prioritizing volume over the harder, slower work of mining my own actual experience for every single piece.
Sitting With the Realization
It was low value, and admitting that plainly is a specific kind of uncomfortable, because it means acknowledging that a strategy I trusted and followed for months was quietly working against the exact thing that makes this project mine. Nobody sets out to build something generic. I set out to build something built entirely on a real, unusual story — no coding background, no budget, an AI-powered build from scratch — and somewhere in optimizing for approval metrics, some of that specificity got flattened out of the actual articles.
There's a particular frustration in realizing the tool that built your entire platform also, without any malicious intent, nudged your content strategy toward exactly the kind of blandness that gets flagged as low value. Not because AI failed at its job. Because I asked it the wrong question — "how do I get approved" instead of "how do I tell this story better than anyone else could" — and it answered the question I actually asked, competently and thoroughly, straight into a wall.
What I'm Doing Differently Now
Now I need to do articles with real value, and that means changing what I'm asking AI to help with, not whether I use it at all. Instead of asking for volume and structural optimization first, the process starts with the personal detail — my actual reaction, my actual mistake, my actual number, my actual moment of doubt or clarity — and AI's job shifts to helping me build the structure, the hook, the pacing around that real material, rather than generating the substance itself.
It's a small shift in sequence, but it changes the entire character of what gets published. An article built AI-first, optimized for approval metrics, tends to read like competent advice. An article built personal-story-first, with AI helping shape and sharpen it afterward, reads like something only I could have written — which is the entire point of a blog that's supposed to be one person's real, unfiltered account of building something from nothing.
AI didn't fail me here. I failed to ask it for the right thing. That's a fixable mistake, and fixing it is exactly what every article in this batch, including this one, is trying to do differently.
Written by Ayoub Zinani, Founder of Fikrago
Explore more: Fikrago Tools · Digital Market · Products
Want help using AI the right way in your own content strategy? Reach out to Ayoub — real lessons from real mistakes.
Image Prompt: A person reviewing two contrasting drafts side by side on a laptop screen, one generic-looking and one personal and handwritten-feeling, warm reflective lighting, thoughtful editing atmosphere, no text overlays, no logos.
(Reminder: add descriptive alt text to this image when uploading to Blogger — e.g. "Comparing generic AI-generated content to a personal, story-driven article.")