A mid-size B2B SaaS company had spent eight months publishing AI-generated content. Traffic was flat, rankings were dropping, and leads from organic search had nearly dried up. Here's exactly what we changed, and what the data showed six months later.
A note on confidentiality: "Acme Co" is a fictitious name used to protect the identity of the real client, who requested anonymity. All other details, including the industry, the content problems identified, and every metric in this case study, are real and drawn from Google Analytics, Google Search Console, and the client's CRM. Numbers have been slightly rounded for readability.
In early 2025, I was contacted by the marketing director at Acme Co, a B2B SaaS company selling compliance and workflow automation tools to mid-market financial services firms. They had a problem that I hear constantly from companies in their position: they had been publishing content consistently for eight months, but nothing was working.
They'd hired a content agency that used AI to produce two to three blog posts per week. The posts were grammatically correct. They covered relevant topics. They were published on schedule. And they were completely invisible, to Google, to their target buyers, and apparently to anyone who might have actually found them useful.
"We were publishing more content than we ever had before. Our traffic was lower than it was two years ago. I didn't understand what was happening."
When I audited the site, the problem was immediately clear. The content looked like it had been generated by prompting an AI with their service category and hitting publish. Generic introductions. Hollow conclusions. Surface-level advice that any competitor could have written. No original insight. No specific examples. No voice. Nothing that would make a financial services compliance officer think "this company actually understands my world."
Worse, some of the AI content had introduced subtle factual errors about compliance regulations, the kind of thing a non-expert wouldn't catch, but that would immediately destroy credibility with their actual buyers.
Before touching anything, I pulled 12 months of data from Google Analytics and Google Search Console to establish a clear baseline. Here's what we were working with:
Despite consistent publishing (2-3 posts/week), organic traffic declined steadily over 8 months of AI-generated content.
Only 4 keywords ranking on page 1. The majority of indexed content sat beyond position 50 — effectively invisible.
My content audit covered every piece of content published in the previous 12 months, 94 articles in total. Here's what I found:
The first priority was stopping the bleeding. We immediately unpublished 49 of the 94 AI articles, specifically the ones with no salvageable content, factual errors, or that were cannibalizing each other's keywords. This sounds counterintuitive, but thin content actively hurts your domain authority. Less is more when the alternative is low-quality filler.
For the remaining 45 articles, I conducted two-hour interviews with both subject matter experts on their team. Every piece of content that went forward needed at least one original insight, one specific example, or one data point that didn't exist anywhere else on the internet. That became the non-negotiable standard.
I also rebuilt their internal linking structure from scratch, connecting related articles into topic clusters and identifying three pillar pages that would anchor their most important keyword categories.
With the foundation in place, we shifted to net new content, but at a dramatically different quality level. Instead of 2-3 posts per week, we published 4-6 pieces per month. Each piece was 1,500-3,000 words, based on primary research or expert interviews, and written specifically for one of three buyer personas we had defined in month one.
I also implemented a GEO strategy, structuring content with clear question-and-answer sections, citable statistics, and authoritative external links that AI tools like ChatGPT and Gemini could extract and reference. Within 60 days, the client's brand began appearing in Perplexity AI responses for two of their target queries.
In the final phase we focused on distribution, repurposing top-performing articles into LinkedIn posts, using the email list to drive return traffic, and building three offsite placements on industry publications that generated quality backlinks. We also added a case studies section to their site (which previously had none) and published two anonymized client success stories.
Here's what the data showed at the 6-month mark compared to our baseline:
Traffic began recovering in month 2 after removing thin content, then grew consistently through months 3-6 as new high-quality content indexed.
Page 1 keywords increased from 4 to 31. Top 3 positions (which capture the majority of clicks) grew from 0 to 9 keywords.
Organic leads grew from 3/month to 24/month, an 8x increase. At the client's average deal value, this represented significant pipeline growth.
| Metric | Before | After (6 months) | Change |
|---|---|---|---|
| Monthly organic visitors | 1,240 | 3,870 | +212% |
| Page 1 keyword rankings | 4 | 31 | +675% |
| Top 3 keyword rankings | 0 | 9 | New |
| Organic click-through rate | 0.8% | 3.4% | +325% |
| Monthly organic leads | 3 | 24 | +700% |
| Brand cited in AI tools | 0 | 3 platforms | New |
| Average session duration | 1:12 | 3:47 | +214% |
| Bounce rate | 78% | 49% | −29pts |
The traffic and ranking numbers are satisfying, but the number that matters most to the client is leads. Going from 3 organic leads per month to 24, at their average deal value of $18,000-$30,000 per annual contract, meaning organic search went from a negligible channel to one of their top two revenue drivers within six months.
The GEO results were arguably the most forward-looking win. Three AI platforms now reference Acme Co when users ask questions about compliance workflow automation, unprompted, without any paid placement. That's brand authority that compounds over time and doesn't require ongoing ad spend.
"I spent eight months watching AI content do nothing. In six months of working with Jennifer, organic search became our second-best lead source. I wish I'd made this call sooner."
AI content isn't inherently bad. The problem is using it as a substitute for expertise rather than a tool to support it. The companies winning at content marketing right now are the ones publishing fewer pieces with more depth, more original insight, and more genuine usefulness to their actual buyers, not the ones publishing the most.
If your blog traffic is flat or declining despite consistent publishing, if your content looks and sounds like every other company in your space, or if you're not showing up in AI search results, the problem is almost certainly quality, not quantity.
The fix isn't complicated. It's just harder than pressing a button.
If your blog isn't ranking, your AI content isn't converting, or your organic leads have dried up, let's talk. I offer a free content audit with no strings attached.
Get Your Free Content Audit