Grounded AEO
Verified business facts that AI answers can cite
Start new thread

Grounded AEO - Verified business facts that AI answers can cite

byβ€’
AI answers now decide who gets found. Grounded AEO audits what ChatGPT, Perplexity & Google AI can verify about your business, then publishes your owner-approved facts as a signed, machine-readable record (API, llms.txt, MCP, DNS) with weekly citation checks. The audit is free, no account needed, and the score can't be bought.

Add a comment

Replies

Best
Maker
πŸ“Œ
Hey Product Hunt πŸ‘‹ I built Grounded AEO after watching AI answers get businesses wrong over and over β€” old addresses, dead phone numbers, or just... nothing. Every tool I found monitors what AI says about you. Nothing fixed the supply side: what the engines actually have available to cite. So this does three things: 1. A free 30-second audit shows what ChatGPT/Perplexity/Google AI can read, verify, and trust about your domain (no account needed) 2. You review and approve your extracted facts, and they're published machine-readable everywhere engines look β€” API, llms.txt, a public grounding page, and an MCP endpoint agents can query directly 3. Records are Ed25519-signed and anchored by an open DNS standard (v=AEO1, spec at aeorecord.org β€” anyone can implement it free) Honesty is the whole angle: the audit score can't be bought, and nothing promises "guaranteed citations" β€” no honest tool can. It measures weekly whether AI actually mentions you, which is the number that matters. Would love to hear how AI answers treat YOUR business β€” run the free audit and tell me what it got wrong.

Ran my site through the audit and was surprised how many little details ChatGPT and Perplexity had wrong. Love that the score can't be bought and the fixes publish as a signed record, feels like the right approach for this AI search era.

Maker

Β Thank you for actually running it β€” that's exactly the moment I built this for. Those small wrong details are what the engines quietly cross-check, so fixing them at the source is most of the game. If anything in the audit felt confusing, tell me β€” plain English is half the product.

Really like the signed record approach and the no-account-needed audit, that lowers the bar a lot. One thing that would help me though is a way to see exactly which line or claim in the audit actually moved my score week over week, right now I get a number but not a clear "you added hours here, lost review count there" diff.

Maker

Β Really appreciate this β€” you've described the gap exactly. Today you get the per-check breakdown (what passed, what failed, and why), but not a clean week-over-week "this exact claim moved your score" diff. The checks already keep history over time, so it's very buildable β€” it just went on the list. Sharpest feedback of the day, thank you.

Ran the free audit on a side project and the result was actually useful, it flagged a couple of third party listings I had no idea were wrong. Having an llms.txt and a signed record shipped out the same week feels like a low effort way to stay honest with the AI crawlers going forward.

Maker

Β Thanks for actually running it β€” those wrong third-party listings are exactly the thing that quietly poisons what the AI crawlers repeat back, so catching them early is the whole point.

If you don't mind me asking, which route are you leaning for the implementation β€” hand-curating the llms.txt and record yourself, or the prepared AI setup-prompt? I built that prompt afterward to fill the gap for the less-experienced devs (paste it into any AI and it walks the whole install), but it's meant to guide the experienced folks just as well. Genuinely curious which fits how you work.