Your customer asked. Your website wasn't in the answer.
Google hands your customer ten links to choose from. An AI hands them three names and a sentence about each. To be one of those three, an assistant has to be able to read you and say something specific about you — and for most websites it can do neither. Here is exactly where it stops.
Watch what an assistant actually does
🔎
A real question, asked of an AI assistant.
✦
Which websites do I know about that I could use to answer this?
solarworks.com.au
✓ fetched
brightspark.com.au
✓ fetched
helios-energy.com.au
✓ fetched
yourbusiness.com.au
✕ robots.txt says no
If AI's crawler can't fetch your site, you never make this list — and everything after this happens without you in it.
You're on the list. Agoria publishes you on a surface that welcomes the AI crawlers by name — and an assistant connected to Agoria doesn't need to crawl at all.
✦
Now let me read them and keep what's useful.
What I've gathered so far
✓solarworks.com.au — CEC-accredited, Newcastle & Hunter, 6.6 kW from $4,990
✕yourbusiness.com.au — page came back empty, content needs JavaScript
If your words only appear after JavaScript runs, nothing of yours goes into that context — the crawler received an empty shell. Your site looks fine to you because your browser runs JavaScript. The crawler doesn't.
Your facts are in the context. Every Agoria surface is rendered on the server, so the words are in the HTML before anything runs — and repeated as structured data.
✦
The question was about the best installers. Which of these actually gives me something to say?
Facts I can use in an answer
✓Accreditation, written as text — solarworks, brightspark
✓A real price to quote — solarworks: from $4,990
✓Suburbs actually named — solarworks, helios
✕yourbusiness.com.au — “contact us for a quote”, accreditation is a badge image
If the fact the question turns on isn't written as text, you can't be used in the answer — even when the crawler read your page perfectly. A badge is a picture. “Request a quote” cannot be quoted.
You're quotable. Agoria turns each of those into a typed field and publishes it as readable text and structured data — so there's something specific to say about you.
✦
I'm about to recommend these to someone. Which can I actually stand behind?
solarworks.com.au
✓ 412 reviews · licence on a register
brightspark.com.au
✓ 189 reviews · industry body listed
yourbusiness.com.au
✕ “fully licensed” — nothing outside your site confirms it
If nothing outside your own website backs you up, you're the risky choice — and an assistant recommending you is putting its own credibility behind you.
You check out. Agoria verifies credentials against the real registers rather than taking your word, and publishes your reviews as structured data.
✦
Here's my answer.
Based on what I can read, three worth contacting:
SolarWorks (CEC-accredited, from $4,990 in Newcastle),
Brightspark (SAA-accredited, 10-year workmanship) and
Helios Energy (Fronius & Tesla, handles the STC paperwork).
Your customer never knew you were an option. You weren't beaten on price or quality — you were absent from every step that decided the answer.
You're in the answer, described in the assistant's own words, to someone who was already asking.
The first two steps decide whether you're a candidate at all. The last two decide whether you're chosen. Most websites never get past the second.
1
🔍 Can AI find you?
Whether the assistant's crawler is allowed to fetch your pages at all.
Most robots.txt files were written years ago to keep 'bots' out. The AI crawlers got caught in the same net — usually by a developer who left long before ChatGPT existed. Nobody decided this; it just happened.
Without AgoriaA site nobody can read
# robots.txt
User-agent: *
Disallow: /
# every AI crawler is refused here
What Agoria does: Agoria publishes your business on a domain that welcomes the AI crawlers by name — and the connector skips crawling altogether: an assistant connected to Agoria reads your record directly, the moment you save it.
What it can't do: Agoria cannot edit your own robots.txt. It can only give you a surface that was never blocked — and tell you exactly which line on your site to change.
2
📖 Can AI read you?
Whether your content exists in the HTML, or only appears after JavaScript runs.
As at 2026 the major AI crawlers do not execute JavaScript — they read the raw HTML and move on. (Google's Gemini is the exception, because it borrows Googlebot's renderer.) So a site built as an app hands the crawler an empty shell. This is the one owners refuse to believe, because the page looks perfect in their own browser. Their browser runs JavaScript. The crawler does not.
Without AgoriaWhat the crawler receives
<div id="root"></div>
<script src="/app.js"></script>
<!-- 0 words of readable content -->
What Agoria does: Every Agoria surface is rendered on the server — the words are in the HTML before anything runs — and the same facts are emitted again as schema.org JSON-LD, which is the format assistants parse most reliably.
3
💬 Can AI answer with you?
Whether the specific fact the question turns on is written down as text.
Nobody asks "who is Acme". They ask who is accredited, who covers their postcode, what it costs. If the price is in a PDF, the licence is a logo in your footer and the service area is implied by your address — none of those are facts a machine can use. A badge is a picture. An assistant cannot look at a picture of a logo and conclude you are accredited.
Without AgoriaFacts a machine can't use
Pricing: "Contact us for a quote"
Accredited: [ badge.png ]
Service area: (implied by address)
Hours: [ hours-table.jpg ]
→
With AgoriaLive at /b/demo-foil-and-fold/llms.txt
Services (as text, and as JSON-LD):
"Colour consultation"
"Half-head foils" — 210
"Full balayage" — 320
Area served: Fitzroy
Rating: 3 reviews on file
What Agoria does: Agoria turns each of those into a typed field, then publishes it twice — as readable text and as structured data. It also shows you the questions customers actually ask in your category, and which of them your record still cannot answer.
What it can't do: llms.txt is published too, but it is not the fix: our own rubric scores it zero, because almost no assistant reads it yet. Gate 3 opens on text and JSON-LD.
4
🛡️ Can AI trust you?
Whether anything outside your own website confirms you are a real, credentialled business.
An assistant recommending you is putting its own credibility behind you. A business that exists only on its own website — no reviews it can read, no link to a register, no third party mentioning it — is a risk it does not need to take when a competitor is easier to stand behind.
Without AgoriaUnverifiable claims
"Fully licensed and insured"
"Award-winning"
"Trusted by thousands"
<!-- nothing links to anything -->
→
With AgoriaChecked, and reported either way
Rating: 4.9/5 (3 reviews on file) → Review schema
Credentials → checked against the register
ABN → "check digit doesn't add up" (said out loud)
What Agoria does: Agoria checks credentials against the real registers rather than taking your word, publishes your reviews as structured data, and links your record to the profiles that prove you exist.
What it can't do: It reports failures too. On this very demo record the ABN did not pass its check digit, and the public file says so — because a trust signal you can fake is not a trust signal.
Which gate are you stuck at?
The scan checks all four against your real website and tells you which one stops you — usually in about a minute, with no account.