← Back to HomeBack to Blog List
Your traffic is up but AI engines aren't citing you — here's how I found out why

Your traffic is up but AI engines aren't citing you — here's how I found out why

📌 Key Takeaway:

Your traffic is up but AI engines aren't citing you. Here's why topical authority beats domain authority, and how to actually measure it.

Last Tuesday I ran a quick audit on our top 20 blog posts. Organic traffic up 23% month-over-month. Great, right? Then I checked how many of those posts were being cited by Perplexity, ChatGPT, or Google's AI Overviews for our target queries. Three. Out of twenty.

That gap between traditional SEO wins and actual AI visibility is what keeps me up at night. And I think most of us are measuring the wrong thing.

The citation problem nobody's talking about

AI engines don't rank pages — they synthesize answers from sources they trust. That trust is built differently than PageRank. I've been tracking this for six months now across three client sites, and the pattern is consistent: high-DA pages with strong backlink profiles often get completely ignored by generative engines.

Why? Because these engines prioritize:

> Source authority within a specific topical cluster, not domain-wide authority.

A post about "Kubernetes pod autoscaling" on a devops blog with DR 45 will outperform a similar post on a general tech site with DR 72 — if the devops site has 15 other posts covering related Kubernetes topics. The engine sees a topical authority signal that the generalist site doesn't provide.

How I actually measure this (the spreadsheet method)

I built a simple tracker in Notion that my team updates weekly. Here's what we monitor:

1. Citation frequency — How often does a specific URL appear in AI-generated answers for our target queries? We check 10 queries per week manually.

2. Contextual positioning — When cited, are we the primary source or a secondary mention? This matters more than you'd think.

3. Competitive displacement — Which sources are replacing us when we drop out of citations?

The numbers are humbling. One client went from being cited in 40% of AI answers for "cloud cost optimization" to 12% in eight weeks. Their organic rankings didn't budge. The content was still there — it just stopped being *trusted* by the engines.

Turns out they'd published three posts on "serverless architecture" in that same period. The engines started seeing them as a generalist cloud blog, not a cost optimization specialist. Topical drift is real and it's silent.

The fix that worked (and the one that didn't)

What didn't work: Adding more internal links. We tried beefing up the internal link structure between cost optimization posts. Zero impact on citation rates over four weeks. What worked: Content pruning and consolidation. We:
  • Merged 4 thin posts (under 800 words) on related topics into 2 comprehensive guides
  • Added explicit "source" sections with data attribution
  • Removed 2 off-topic posts that were diluting our topical signal
  • Citation rate bounced back to 35% within six weeks. No new backlinks. No technical SEO changes. Just cleaner topical signaling.

    The tool gap nobody's filled yet

    Here's the frustrating part: we have amazing tools for traditional SEO. But for GEO? You're mostly doing manual checks or using GEO Audit Tool to get a baseline read on where you stand. It's not perfect — the sample sizes are small — but it's better than guessing.

    I've also been experimenting with the AI Gravity Checker to see which of our pages have the structural signals that engines seem to favor. The correlation isn't perfect, but pages scoring above 70 on their scale get cited roughly 3x more often than those below 40. Worth looking into if you're trying to prioritize what to fix first.

    What I'm doing differently now

    I stopped treating GEO as a separate channel. It's not "SEO plus AI stuff." It's a completely different trust model.

    My new rule: if a post can't answer the question "Why should an AI engine cite this specific page instead of the 50 others on this topic?" — it doesn't get published.

    That means:

  • Every post needs a unique angle or dataset
  • We attribute primary sources explicitly
  • We maintain tight topical clusters (no more "and another thing" posts)
  • We prune ruthlessly when content drifts
  • The GEO vs SEO breakdown I read last month put it well: SEO is about being found. GEO is about being *chosen* as a source. Different game.

    The uncomfortable truth

    Most of our content strategies are still optimized for 2019. We're chasing rankings while AI engines are building knowledge graphs that bypass rankings entirely.

    I'm not saying ditch SEO. I'm saying if you're not actively measuring whether AI engines are citing your content, you're flying blind. The traffic might look fine right now. But in 12 months? I think we'll see a much sharper divide between sites that engines trust and sites that just rank well.

    Start tracking citations manually this week. Pick 5 queries where you want to be the source. Check Perplexity, ChatGPT, and Google AI Overviews. Note where you appear and where you don't. That baseline will tell you more about your GEO health than any ranking report.

    Want Better SEO Results?

    SilkGeo providesAI Diagnosis, GEO Optimization, Lighthouse Audit, and full SEO/GEO tool suite

    Use SilkGeo for free