findloc.ai
GEO · 5-minute explainer

When someone asks ChatGPT “best café in Brooklyn”, whose name comes up?

That answer is decided by 5-7 signals AI looks at — and most businesses get less than half of them right. Here’s what actually works, what doesn’t, and how to tell when an agency is overcharging you.

What is GEO, in plain English

When ChatGPT (or Perplexity, Claude, Google AI) answers a question, it doesn’t just look at one website. It checks multiple signals and picks businesses that show up consistently across them. For a hotel in Boston, the AI is checking:

SourceWhat AI looks at
The hotel’s own websiteClear <title>, schema.org Hotel structured data
Google Business ProfileReviews, photos, hours, address
TripAdvisor / Booking.comRatings and reviews
Wikipedia / travel blogsThird-party mentions
Local news / tourism mediaEditorial citations
/llms.txt fileFew sites have one — AI prefers ones that do

A business that “wins” in AI answers has these 5-6 sources all aligned — same name, address, phone; consistent description; structured data on the website. That’s GEO.

What actually moves the needle

Ranked honestly by how much each action actually changes AI behaviour. Stars = AI impact.

  • 1.Google Business Profile, complete & updated
    ⭐⭐⭐⭐⭐Easy · 1 hour
  • 2.Schema.org LocalBusiness markup on your website
    ⭐⭐⭐⭐⭐Medium · needs a developer
  • 3.Mentions on 5-10 real third-party sites (news, Wikipedia, TripAdvisor)
    ⭐⭐⭐⭐⭐Hard · takes time
  • 4.NAP (Name, Address, Phone) consistent across all platforms
    ⭐⭐⭐⭐Medium · tedious but doable
  • 5.FAQ section on your site (“What are your hours?” etc)
    ⭐⭐⭐⭐Easy · 1 hour
  • 6.Working sitemap.xml + robots.txt
    ⭐⭐⭐Easy
  • 7./llms.txt file (the new “robots.txt for AI”)
    ⭐⭐⭐Easy

Honest fact: 80% of small businesses haven’t even done step 1. If you’re paying an agency for “AI SEO” and they haven’t finished your Google Business Profile, you’re being overcharged.

Why agencies charge $5,000+ for $500 of work

Nothing against agencies — we just believe in transparent pricing. A typical $4,000–7,000/year “AI optimisation” package usually breaks down like this:

What they doReal hoursActual value
Add schema.org markup1-2hHigh
Submit to 50-100 directories~10hLow (most directories are dead)
Write 5-10 generic blog posts5-15hLow (AI ignores self-promotion)
Send monthly dashboard reportsautomatedLow (hard to verify)
Wikipedia / press outreach0-5hHigh (but rarely actually done)

Real cost of delivery: ~$1,000–1,800. They charge $4,000–7,000 because most clients can’t evaluate what they got. We do the same work, transparently, on Pro for $588/year.

Convinced? See pricing.

Free to claim. Pay only when you want real Google reviews synced + weekly proof of which AI engines are citing you.

See pricing →