15 years in agencies and I've seen every 'new SEO'. What makes GEO different — or not?
Not here to be a buzzkill, genuinely asking. I've been through local SEO, voice search, featured snippets, SGE panic, AIO panic. Each one had a cottage industry of consultants selling urgency.
Here's my honest read on GEO in May 2026:
The real part: AI search is genuinely influencing purchase discovery now. Tinuiti's Q1 numbers are not hype. I've seen it in client pipelines. This part is real.
The grift layer: Agencies charging $5k-$10k/month for 'AI visibility optimization' and delivering llms.txt files and schema audits. That work is real too, but it costs $500 in a tool and a junior's time, not a retainer.
The genuinely new stuff vs. repackaged SEO:
- New: per-engine retrieval strategies differ. Claude vs ChatGPT vs Perplexity have different trust hierarchies. That's actually novel.
- New: training-data lag means you can do everything right today and see nothing for 6-8 weeks. SEO had crawl lag but nothing this long.
- Repackaged: comparison pages, review velocity, schema markup. All of this worked in 2015.
So what am I missing? What's the thing that would make me say 'okay, this is fundamentally different from every other SEO pivot'?
4 replies
- Jess Wright
The training-data lag point is what separates GEO from every previous SEO wave for me. Doing everything right today and seeing nothing for 6-8 weeks is psychologically brutal in a way that SEO never was. At least with SEO you'd see crawl confirmation in hours. The delayed feedback loop makes it really hard to know if you're on the right track.
- Petr VlčekFounder
Petr here — I think the thing that makes GEO genuinely different is the entity model. Classic SEO had keyword-page matching. GEO runs on entity graphs — the model has an opinion about what your company *is*, not just what pages you have. Changing that opinion requires changing external evidence, not just your own pages. That's a fundamentally different feedback loop.
- Milan Novák
Counterpoint on the repackaged SEO stuff: comparison pages and review velocity 'worked in 2015' but they work for a different reason now. In 2015 they got you Google organic rankings. In 2026 they influence AI entity representation. Same tactic, different mechanism. Worth understanding the new mechanism even if the action looks the same.
The 'different trust hub per engine' point is what actually changed how I work. Once I stopped thinking of 'AI optimization' as one thing and started thinking about Claude optimization vs ChatGPT optimization vs Perplexity optimization as separate problems, my approach got way more targeted.