AI & Automation
Personas
SaaS & Startup
Time to ROI
Medium-term (3-6 months)
So here's the thing that happened six months ago that changed how I think about SEO forever. I was working with a B2B SaaS client, and we'd just finished this massive content overhaul - you know, the usual stuff. Blog posts, case studies, the whole nine yards.
Everything looked great in our traditional SEO tools. Ahrefs was happy, SEMrush was showing green arrows, and our Google rankings were climbing. But then something weird started happening. Our prospects were mentioning specific details from our content during sales calls, but they weren't coming through our tracked organic channels.
That's when I realized we had a blind spot the size of a crater. People were finding our content through ChatGPT, Claude, and other AI tools, but we had zero visibility into this traffic. It was like having a store where half your customers were entering through a secret door you didn't even know existed.
Here's what you'll learn from my experience figuring this out:
Why traditional SEO tracking misses AI-generated traffic completely
The manual methods I developed to track mentions in ChatGPT and Claude
How to optimize your content for AI recommendation engines
The surprising metrics that actually matter in the AI search era
Tools and workflows that give you visibility into this "dark traffic"
The reality? We're in the middle of the biggest shift in how people find content since Google launched. And most businesses are flying blind. Let me show you what I learned when I had to figure this out the hard way.
Industry Reality
What traditional SEO tracking completely misses
Right now, most companies are still living in the Google-first world when it comes to tracking their content performance. I get it - it's what we've all been doing for the past two decades. Open up Google Analytics, check your keyword rankings in Ahrefs, see which pages are getting organic traffic, and call it a day.
The industry standard approach goes something like this: track keyword positions, monitor organic click-through rates, measure time on page, and analyze conversion rates from search traffic. Most SEO professionals are obsessing over featured snippets, local pack rankings, and whether their content shows up on page one for their target keywords.
Here's the problem with this entire approach - it's based on the assumption that people are still primarily discovering content through traditional search engines. But that's not what's happening anymore, especially in the B2B space.
According to every marketing blog you'll read, you should be focusing on optimizing for search intent, building topic clusters, and creating content that ranks well in Google. All of this advice assumes that Google is still the primary way people discover and consume business content. But what happens when your ideal customers are asking ChatGPT "What's the best inventory management solution for Shopify?" instead of googling it?
The conventional wisdom completely breaks down because traditional SEO tools can't track AI-generated recommendations. Google Analytics will never show you that someone discovered your product through a ChatGPT conversation. Ahrefs can't tell you that Claude recommended your blog post to someone researching your industry.
This creates a massive blind spot. You might think your content strategy isn't working because your traditional metrics are flat, when in reality, AI tools are driving significant qualified traffic that you just can't see. It's like judging a store's success by only counting people who walk through the front door, while ignoring everyone who enters through the side entrance.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
OK, so here's exactly what happened. I was working with this B2B SaaS client - they had a pretty solid content operation going. We'd spent months optimizing their blog, case studies, and product pages. Traditional metrics looked decent, but something felt off during their sales calls.
The prospects knew way too much about their specific use cases and features. They were asking questions that suggested they'd read our most detailed blog posts, but when I checked Google Analytics, the referral sources didn't add up. We were seeing qualified leads who seemed extremely informed, but our content attribution was basically showing nothing.
I started digging deeper during the sales process. Instead of just tracking where leads came from, I began asking prospects directly during discovery calls: "How did you first learn about us?" and "What research did you do before reaching out?"
That's when the pattern emerged. About 40% of our most qualified prospects mentioned something like: "I asked ChatGPT about inventory management solutions" or "I was researching this problem and found your approach through Claude." Some even said they'd specifically asked AI tools to compare different platforms, and our content kept coming up in the responses.
This was traffic that traditional analytics couldn't see. These weren't people clicking through from Google searches. They weren't finding us through social media or referrals. They were having conversations with AI tools, getting recommendations for our content, then typing our URL directly into their browser or searching for our brand name.
From an analytics perspective, this showed up as "direct traffic" or "branded search," which completely masked the real source. We had this entire acquisition channel that was invisible to our normal tracking methods. It was like discovering that half our customers were entering our store through a hidden entrance we didn't know existed.
The crazy part? When I started manually testing this by asking ChatGPT and Claude questions related to our client's industry, our content was indeed showing up in the responses. But we had no systematic way to track this, no metrics to optimize for, and no understanding of which content was performing best in AI recommendations.
Here's my playbook
What I ended up doing and the results.
Right, so once I realized we had this massive blind spot, I had to figure out how to actually track performance in AI tools. There's no Google Analytics for ChatGPT mentions, so I had to get creative and build my own system.
Step 1: Manual AI Content Auditing
I started by creating a weekly audit process. Every Monday, I'd spend 2 hours systematically testing different queries in ChatGPT, Claude, and Perplexity. I'd ask questions our prospects would typically ask: "What's the best solution for [specific problem]?" "How do companies handle [use case]?" "Compare different approaches to [industry challenge]."
The key was thinking like our target customers, not like marketers. I'd ask follow-up questions, request specific recommendations, and even ask for pros and cons comparisons. Then I'd document every mention of our brand, content, or concepts in a spreadsheet with the query used, which AI tool, and how prominently we appeared.
Step 2: Content Optimization for AI Recommendation
Based on what I learned from the manual audits, I started optimizing our content specifically for AI tools. This meant restructuring articles to be more question-and-answer focused, adding clear problem-solution frameworks, and including specific use cases with concrete examples.
I discovered that AI tools love content that's structured like: "For [specific situation], the best approach is [solution] because [clear reasoning]." They also tend to recommend content that includes real metrics, specific examples, and step-by-step processes rather than generic advice.
Step 3: Tracking "Dark Direct" Traffic
Since AI-sourced traffic often shows up as direct visits or branded searches, I created a system to identify this hidden channel. I set up UTM tracking for all our content URLs and started including those tracked links in our email signatures, social profiles, and anywhere we could control the entry point.
The difference between UTM-tracked direct traffic and untracked direct traffic gave us a proxy for AI-influenced visits. I also set up custom events in Google Analytics to track when people viewed multiple pieces of content in a single session - a behavior pattern that was more common with AI-sourced traffic.
Step 4: Sales Process Integration
I worked with the sales team to add specific discovery questions to their qualification process. Instead of just asking "How did you hear about us?" they started asking: "What research process led you to reach out?" and "What specific information convinced you we might be a good fit?"
This qualitative data became crucial for understanding not just that AI tools were driving traffic, but which specific content and messaging resonated most with prospects who discovered us through AI recommendations.
Step 5: Competitor AI Presence Monitoring
I extended the manual audit process to include competitive intelligence. By asking AI tools about our competitors and industry alternatives, I could see where we had gaps in coverage and identify opportunities to improve our AI visibility in specific topic areas.
This also helped me understand the broader landscape of how AI tools were recommending solutions in our industry, which informed our content strategy and positioning.
Manual Auditing
Test 20+ industry queries weekly across ChatGPT, Claude, and Perplexity to track mention frequency and prominence
Content Structure
Optimize articles with clear problem-solution frameworks and specific examples that AI tools prefer to recommend
Dark Traffic Tracking
Use UTM parameters and behavioral analytics to identify AI-influenced visits that appear as direct traffic
Sales Integration
Add discovery questions to qualification process to capture AI-sourced lead attribution qualitatively
After three months of implementing this system, the results were pretty eye-opening. We discovered that roughly 35% of our qualified leads were influenced by AI tool recommendations, but this traffic was completely invisible in our traditional analytics.
The manual auditing process revealed that our content was mentioned in AI responses about 60% of the time when prospects asked questions related to our core use cases. More importantly, we were often the first or second recommendation in longer AI responses, which correlated with higher-quality lead generation.
From a business impact perspective, leads that came through AI discovery had a 40% higher close rate compared to traditional search traffic. This made sense - these prospects had essentially received a personalized research session before reaching out, so they were more informed and better qualified.
The sales team reported that AI-sourced prospects asked more sophisticated questions during discovery calls and had a clearer understanding of our differentiators. They were less price-sensitive and more focused on implementation and results, which made the sales process more efficient.
Unexpectedly, we also found that content optimized for AI recommendations performed better in traditional search over time. The clear structure and specific examples that AI tools preferred also seemed to improve our Google rankings and featured snippet appearances.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
AI discovery is becoming the new "dark social" of B2B marketing. Just like social media traffic that doesn't get properly attributed, AI-influenced traffic is a growing blind spot that most companies haven't acknowledged yet. You can't optimize what you can't measure.
Traditional SEO metrics are becoming less reliable indicators of content performance. We had articles that ranked poorly in Google but were frequently recommended by AI tools to qualified prospects. Conversely, some high-ranking content never appeared in AI responses.
AI tools prefer authoritative, specific content over generic advice. The content that performed best in AI recommendations included real metrics, specific use cases, and clear problem-solution frameworks. Generic "best practices" articles rarely got mentioned.
Manual tracking is currently the only reliable method. There are no automated tools yet that can track your presence in AI recommendations across different platforms. You have to invest time in systematic manual auditing if you want visibility into this channel.
Sales process integration is crucial for measurement. Without changing how you qualify leads and ask discovery questions, you'll never understand the true impact of AI-influenced traffic on your business results.
Content velocity matters more than content volume. AI tools seem to favor more recent content when making recommendations. Regularly updating and refreshing existing content improved our mention frequency more than creating new articles.
This trend is accelerating, not slowing down. As AI tools become more integrated into people's research workflows, traditional search behavior is shifting. Companies that figure out AI optimization early will have a significant advantage.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS companies specifically:
Focus on use-case specific content that answers "How do I solve [specific problem] with [your type of solution]?"
Include integration examples and API documentation that AI tools can reference
Structure comparison content that positions your solution against alternatives
Track branded search increases as a proxy for AI-influenced discovery
For your Ecommerce store
For ecommerce stores:
Optimize product descriptions with specific use cases and benefits that AI tools can extract
Create buying guides that AI tools can reference when customers ask for recommendations
Include customer reviews and specific metrics that establish authority
Monitor direct traffic spikes after creating AI-friendly content