You ask AI for "fried chicken", whose website wins? A&W vs KFC vs McDonald's

Ask ChatGPT "where's the best fried chicken near me" and it has to quote some website. Billion-dollar brands assume it'll be them. So I checked.

Background, why I built this

I'd built a GEO audit agent that scores any URL on how ready it is to be cited by AI search. Scoring my own blog was fine, but the real test is famous brands with big budgets. Fried chicken is perfect: everyone knows A&W, KFC, and McDonald's, and everyone types "fried chicken" into something.

What

A head-to-head GEO audit of the three chains' Indonesian websites, A&W (aw.co.id), KFC (kfcku.com), McDonald's (mcdonalds.co.id), scoring each on whether an AI assistant can crawl, parse, and quote it.

Why it matters

AI assistants don't show a menu of ten links, they pick a few sources and quote them. For a category as competitive as fast food, the brand whose website is machine-readable gets recommended; the others simply aren't in the answer, no matter how big the marketing budget.

Who it's for

Brands, marketers, and anyone who assumes "we're famous, of course AI mentions us." Fame is not the same as being citable.

When & where

Run live against the public Indonesian sites. The audit checks each page plus the domain root (robots, llms.txt, sitemap).

How

The same seven weighted GEO checks, AI-bot access, llms.txt, sitemap, structured data, crawlable content, answer structure, meta. Here's where each chain wins or leaks:

What the data showed:

• McDonald's leads, but only "Decent" (59). It's crawlable and has a sitemap, but no llms.txt and no structured data. It's findable, not optimised.
• KFC is "Weak" (37). No sitemap, no schema, thin answer structure. A bot can reach it but struggles to understand it.
• A&W is invisible (0). The bot couldn't even fetch aw.co.id, blocked or JS-locked at the door. If a crawler can't load your page, you don't exist to AI, full stop. (Their US site scored 52, so it's a config problem, not the brand.)
• The whole sector leaves the lane open. Not one of them has an llms.txt or real structured data. For "fried chicken," AI has weak sources to choose from, which means a smaller, well-optimised player could out-cite all three.

The takeaway

Budget buys ads; it doesn't buy citations. A billion-dollar brand with a JS-locked, schema-less site loses the AI answer to a tiny blog that's simply readable. In AI search the question isn't "who's biggest", it's "who can the model actually quote." Right now, in fried chicken, the answer is: barely anyone.

→ Audit your own site free, same engine, instant score + fixes. If billion-dollar chains score in the 30s, your competitors probably do too.

---

Honest limits: a low score means low readiness, not that the brand never appears, AI also pulls from third-party reviews, maps, and aggregators. This audits the brand's own site. Scores are a live snapshot and configs change.