This newsletter is brought to you by Your Content Mart.

Your buyers are asking ChatGPT, Gemini, Claude, and Perplexity questions like "best [category] tool for [use case]" and "what are the best alternatives to [x] for [y]". 

These are money prompts, the AI equivalent of high-intent keywords.

Most B2B SaaS companies have no idea if they show up in those results. Their competitors might.

We built a free AI Search Gap Analysis that tests your visibility across these prompts and shows you exactly where you're missing and how to fix it.

Takes 30 seconds to request. We handle the rest.

[Get Your Free AI Search Gap Analysis →]

Last week, a user of an automated AI content tool shared a thread on X. Traffic had been growing, rankings climbing, and impressions approaching 30,000 per month. Then Google's recent core update hit, and everything tanked. 

His summary: “A simple case of 'it worked until it didn't.”

He wasn't alone. The entire thread was filled with similar stories.

One commenter with 25 years of SEO experience put it bluntly: "Not a SINGLE ‘SEO on Autopilot' tool in history has ever worked long term. The entire ethos is contrary to implementing solid SEO best practices and long-term growth."

Others pointed out that the tool wasn't just creating low-quality content at scale. It had also built a faux backlink exchange network. Users were getting hit from both sides: thin content and unnatural links.

Lily Ray, one of the most respected voices in the SEO industry, weighed in too. Her take: if a popular SEO tool promises a big surge in traffic, expect it to crash within a few months.

But this isn't just anecdotal. We now have data that confirms it at scale.

SE Ranking published an experiment that makes the pattern undeniable.

They published 2,000 AI-generated articles across 20 brand-new websites. No human editing. No internal linking. No images. No link building. Pure AI content, left to perform on its own.

The early numbers looked promising. About 71% of pages got indexed within 36 days. Eight sites ranked for over 1,000 keywords each within the first month, and impressions reached over 122,000 with 240+ clicks.

Then, on February 3, 2025, every single site lost all traction. Rankings gone. Traffic gone. The same pattern from the X thread, replicated across 20 domains.

But SE Ranking also ran a parallel test. They published six AI-assisted articles on their own established blog. Same AI tool, but these articles were reviewed by their team, edited for accuracy, and enhanced with real expertise. Those six articles generated nearly 555,000 impressions and over 2,300 clicks. Three currently rank in the organic top 10, and four appear as sources in Google's AI Overviews.

The difference wasn't the AI. It was the strategy, the human oversight, the domain authority, and the quality standards applied to the content.

One of the commenters in that X thread captured it well: 

That's the core issue. What happens when a blog post goes out of date and needs a refresh? What about when the search intent behind a keyword shifts? How do you rank for competitive terms without link building? These are strategic questions, and no AI tool is answering them for you.

If your AI content is ranking well today, that doesn't mean you're in the clear. 

The SE Ranking experiment shows that early results can look genuinely promising. The 2,000 articles indexed fast, ranked for thousands of keywords, and generated real impressions. The crash came later. The early data made AI-at-scale look like a winning bet. The problem is that the early data was misleading.

So, does this mean AI is useless for content creation? Absolutely not.

Does it mean you need to rethink how you use it? Absolutely yes.

The problem isn't AI itself. The problem is what you feed it.

When you give AI a generic prompt like "Write a blog post about proposal software," you get generic output. The same output that your competitors are generating from the same prompt.

When you give AI unique inputs that nobody else has, the output changes dramatically. Here are four ways to do that.

Reach out to practitioners who live in the space you're writing about.

Say you're a proposal software startup writing about best practices for proposal management. Instead of letting AI summarize what already exists, interview three or four people who manage proposals daily. Ask them specific questions:

Those answers give you insights no other article has, because they came from real conversations rather than recycled web content.

Then use AI to structure the article, weave the quotes in, and draft the narrative around those unique insights.

The result is an article that sounds like it was written by someone who understands the space, not a summary of page one.

The SE Ranking experiment backs this up. Their AI-assisted blog posts performed well because they were reviewed by real experts. The 2,000 pure AI articles on new domains crashed. The "human in the loop" factor isn't optional. It's what separates content that ranks from content that gets wiped out in the next update.

If you have access to data that nobody else can get, you have a massive content advantage.

Imagine you run a proposal software startup. You have thousands of users sending proposals every day. That means you have data on average proposal close rates by industry, the most common proposal length that converts best, what day of the week proposals are most likely to get signed, and how quickly prospects respond on average.

Build a blog post around that data. Something like: "We Analyzed 50,000 Proposals. Here's What Actually Gets Signed."

AI can help you analyze the data, spot patterns, draft the narrative, and create compelling visualizations. But the data itself is your unfair advantage.

And this matters for SEO specifically because original research attracts backlinks. When you publish proprietary data, other blogs, journalists, and industry publications cite it. The SE Ranking AI content experiment, for instance, has garnered 233 backlinks from sites like Search Engine Land, Rank Tracker, and Foundation Marketing.

That's the power of proprietary data. It turns a blog post into a linkable asset.

This one is especially relevant for SaaS companies, and most are sleeping on it.

When your product solves the exact problem you're writing about, show it in action. If you're writing about "how to create a professional proposal," walk the reader through creating one using your tool. Include screenshots. Show the step-by-step process. Highlight specific features that make the workflow easier.

This works for three reasons.

First, it makes your content genuinely useful. The reader doesn't just learn the concept; they see exactly how to execute it.

Second, it creates content that AI cannot replicate. Screenshots, product interfaces, and real walkthroughs are unique to your product. No prompt will generate them.

Third, it naturally positions your product as the solution. You're not writing a sales pitch. You're writing a helpful tutorial that happens to feature your tool. That's bottom-of-funnel content done right.

Now, you might be thinking: "Doesn't this make the content too promotional?" Not if you lead with value. Teach the concept first, then show how your tool makes it easier. The reader gets genuine help either way and can see your product in action.

This is different from sourcing expert interviews for quotes (fix #1). This is about who reviews and edits the final output before it goes live.

AI doesn't know your industry the way a human expert does. It doesn't know that its advice on proposal pricing strategies is outdated, that the "best practice" it recommended actually annoys enterprise buyers, or that a subtle technical error will cost you credibility with experienced readers.

Hire content writers or editors who understand your space. Not generalist freelancers who write about everything from dog food to data security. Find people who have worked in your industry, used tools like yours, or have deep knowledge of your customers' problems.

Even if AI handles first drafts, a domain-expert editor catches what it misses.

If you're a well-funded SaaS company with a DR of 70+, you might survive a traffic crash. Your domain authority gives you a buffer.

But if you're a startup with a DR of 20 or 30, competing against established players, you don't have that cushion. A traffic wipeout like the ones we've been discussing could erase months of progress.

Generic AI content is a race to the bottom, and startups lose that race because bigger players have more pages, more backlinks, and more brand recognition. The way you win is by publishing content built on expert interviews, proprietary data, product walkthroughs, and domain-expert editorial oversight. Content that Google can't find anywhere else.

AI-generated content isn't inherently bad. But generic, automated AI content is a ranking liability waiting to surface. The X thread showed real people experiencing real traffic crashes. The SE Ranking experiment proved it at scale: 2,000 AI articles, 20 websites, all rankings gone within three months.

The fix: stop treating AI as your content strategy. Start treating it as a tool within your content strategy. Feed it expert interviews, proprietary data, product walkthroughs, and domain expertise. That's how you turn AI from a content commodity machine into a competitive advantage.

Got any questions or concerns? Reply to this email and let me know.

To your startup's success,

Shehu AbdulGaniy

Founder, Your Content Mart

P.S: If you learned something new from this issue, invite your colleagues and friends to join The SaaS SEO Insights. Counting on your support 🤝