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'?