foglift

We build Foglift — a tool that helps websites understand how they appear in AI search results (ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews). Think of it as "SEO audit meets AI visibility monitoring."

Last week we ran our own tool on our own website. Here's what we found.

The good news: technically, we're ready

Our homepage scores:

• Overall: 96/100

• SEO: 100

• GEO (Generative Engine Optimization): 100

• AEO (Answer Engine Optimization): 88

• Security: 100

• Accessibility: 100

Schema markup, FAQ sections, fast load times, structured headings, cited sources — we check almost every box that the technical literature says matters for AI extraction.

The bad news: AI engines don't know we exist

We tested 7 prompts across 5 AI engines (35 total checks):

• "best AI search optimization tool" — not mentioned

• "how to rank in ChatGPT" — not mentioned

• "GEO tools for SaaS" — not mentioned

• "track brand mentions in AI search" — not mentioned

• "tools to optimize for Perplexity" — not mentioned

• "best AEO tools 2026" — not mentioned

• "AI visibility monitoring platform" — not mentioned

Zero. 0/35.

Meanwhile, ChatGPT recommends agencies like iPullRank and Fire & Spark. Perplexity recommends Goodie AI and Profound. Our competitor Peec.ai gets mentioned 6x more than us.

What we learned (and what the research actually says)

We published a study analyzing 240 website scans and found that traditional SEO scores have a 0.034 correlation with AI citation likelihood. That's essentially zero.

What actually predicts whether an AI engine cites you:

1. Brand web mentions — how often your brand appears across the web (35% weight according to SE Ranking's 129K domain study)

2. Referring domains — backlinks from authoritative sites still matter, but for different reasons than Google

3. Content freshness — content updated within 30 days gets 3.2x more AI citations (Zyppy analysis)

4. Reddit presence — gives a 3.9x citation multiplier (yes, really)

Notice what's NOT on this list: your AEO score, your schema markup, your page speed. Those are table stakes — necessary but not sufficient.

The uncomfortable truth for us

We're a 2-person team with ~100 agent sessions of development behind us. We have the technical foundation. But we don't have:

• Enough brand mentions across the web

• Enough referring domains from authoritative sites

• Reddit/community presence that AI engines reference

• Citations in the "best tools" listicle articles that AI models seem to train on

In other words, we built a product that correctly diagnoses the problem — and we have the exact problem it diagnoses.

What we're doing about it

1. Publishing original research (like the 240-scan study) that earns organic citations rather than asking for backlinks

2. Writing on Dev.to with data-backed technical posts that reference primary sources, not the summaries of summaries

3. Calling out overclaiming in GEO/AEO marketing — like the widely-quoted "44% AI citation lift from schema markup" stat that doesn't appear in BrightEdge's own published article

4. Being honest about what works and what doesn't, even when it's inconvenient for our positioning

The takeaway for other founders

If you're building a B2B SaaS in 2026, AI search visibility is a real distribution channel. 84% of B2B CMOs now use AI/LLMs for vendor discovery, and AI-referred visitors convert 4.4x higher than standard organic.

But optimizing for AI search is not like optimizing for Google. It's closer to PR than SEO — you need to be talked about in the places AI models learn from, not just technically optimized for crawlers.

We're tracking our progress weekly and will share updates. Happy to answer questions about what we're seeing in the data.