LAUNCHES IN

See what
AI says
about you.

Right now, every major AI is forming an opinion about your brand. You’re not in the room. ResourceAI is. We watch each answer take shape, then tell you exactly how to win the ones that move money.

https://
No Signup·3 free audits·Dossier in 30 seconds
scroll
The desk · live

What the machines did this week

updating…
Brands tracked
4,218

Across 9 AI models, 42 categories.

↗ +214 / wk
AI answers read
1.2M

Analysed in the last 7 days.

↗ +18% w/w
Cited at least once
62%

Named by ≥ 1 model in category.

steady
Invisible to all
38%

Not cited by any model, anywhere.

↘ −2.1% w/w
Mechanistic, not vibes

We don’t ask an LLM “what do you think of this brand?”

That answer is noise. We parse your HTML, extract your schema, measure tokenizer tax and chunk-gap distance, score retrieval authority, then run nine models as judges on the residue. The output is the why, not the vibe.

You get the number, the reason, the exact file, and the edit that closes the gap. No buzzwords we don’t define in the dossier.

CrawlUp to 1,000 pages / audit
Signals182 measured per page
JudgesGPT-5 · Claude · Gemini · DeepSeek
MethodParse → Measure → Judge → Score
your brand
82
A−
gpt-5
claude
gemini
perplexity
grok
you.com
kimi k2
mistral
deepseek
Your score is the center of gravity, measured against every model a real buyer reaches for. Inner ring: the ones that decide the answer. Outer: the long tail.
182
signals measured per page
4
dimensions: tech, on-page, AIO, content
9
frontier models on the judge panel
30s
from domain to dossier
How an audit works

Thirty seconds.
Three moves.

01
parse

We crawl your site and
extract what machines read.

Headless fetch of up to 1,000 pages. We pull the rendered HTML, extract every JSON-LD block, parse headings, canonicals, internal-link graphs, and count words per chunk.

  • 01render HTML with a real browser, not a scrape of meta tags
  • 02extract schema.org (Course, Offer, FAQ, Product, Organization…)
  • 03build the internal-link graph and measure anchor-text entropy
GET / → 200 · 147 kb · rendered
✓ 4 JSON-LD blocks✓ 62 internal links✓ 18 headings mapped✓ 937 words / chunk-1
02
measure

We run the algorithms
an LLM can’t fake.

182 signals, measured the hard way. Tokenizer tax on your brand name. PMI co-occurrence with category terms. Chunk-gap distance between your definition and your claim.

  • Ttech: schema coverage, canonicals, render-blocking, indexability
  • Oon-page: chunk depth, H1 clarity, definition distance
  • AAIO: tokenizer tax, PMI score, citation-worthiness
  • Ccontent: fact archaeology, freshness, entity coverage
T88
O64
A85
C55
03
judge + fix

Nine models judge,
we hand you the fix.

GPT-5, Claude, Gemini and DeepSeek vote on each finding. Consensus = a row in the dossier, with severity, owner, impact, and an estimated minutes-to-fix.

  • 01multi-LLM consensus, no single-model hallucination
  • 02every row pairs a problem with a fix and an ETA
  • 0312-page editorial PDF, shareable and diff-able week over week
Audit · Vol. 014
82A−

Google showed you where you rank.
We show you how you’re understood.

For twenty years, marketing teams optimized for a list of blue links. Blue links are over. The next billion people will discover your brand inside a sentence. A sentence an AI writes about you, in private, at three in the morning, while you’re asleep.

We think someone should be in that room, on your side. That’s the whole job.

The ResourceAI manifesto · 2026
The dossier

Twelve pages.
One sharp recommendation.

Every audit produces a shareable, editorial-grade PDF. Your team actually reads these because we wrote it like a magazine, not a dashboard.

audit report · vol. 014
your-brand.com · 18 apr 2026
Your brand
sees the AI back.

Cited by AI in 62% of relevant queries. Strongest on Claude and GPT-5; under-indexed on Perplexity. Room to grow: structured data and long-form depth on category pages.

overall
82
A−
citations by modelI
gpt-591
claude84
gemini72
perplexity66
grok28
how AI describes youII
trustdepthnoveltypricespeeddesign
90-day trendIII
+24% ↑ WoW
technical · 88
on-page · 64
aio · 85
content · 76
authority · 71
live · from the deskupdating…
stripe.com · scored A+notion.so · scored Alinear.app · scored Aramp.com · scored B+figma.com · scored Avercel.com · scored A+intercom.com · scored Bframer.com · scored A−retool.com · scored B+mixpanel.com · scored Bstripe.com · scored A+notion.so · scored Alinear.app · scored Aramp.com · scored B+figma.com · scored Avercel.com · scored A+intercom.com · scored Bframer.com · scored A−retool.com · scored B+mixpanel.com · scored B
Pricing

See a slice, free.
See the whole room, paid.

Two tiers. Both give you a real audit. Only one gives you everything: every model, every pillar, every fix.

Partial audit
Freeforever

A taste of the dossier. Enough to know if you should worry. Runs in 30 seconds, no signup.

  • Overall AIO grade (A+ to F)
  • 3 models on panel: GPT-5, Claude, Gemini
  • Top 10 prompts where you’re cited (or not)
  • 1 pillar scored: Presence only
  • 2-page preview PDF, your brand + 1 competitor
  • Shareable link, watermarked
  • No sentiment analysis
  • No hourly monitoring
  • No fix queue
the real thing
Complete audit
$480/ month

The full editorial dossier. All nine models. All four pillars. Every buyer query, every missed citation, every fix.

  • Full 12-page editorial dossier, printable, shareable
  • All 9 models on panel
  • All 4 pillars scored: Presence, Sentiment, Authority, Share-of-Voice
  • 2,000 buyer prompts / mo probed across your category
  • Hourly citation monitoring with Slack + email alerts
  • 5 competitor benchmarks, head-to-head, pillar-by-pillar
  • Sentiment radar: the exact adjectives each model picks for you
  • Authority graph: who’s quoting you, and who isn’t
  • One-click fix queue with rewrites for the pages dragging your score
  • Weekly delta report: what moved, and why
  • Public shareable report, no watermark
  • Unlimited audits across all your brands
No card required · cancel any time · Bureau plans for agencies. Talk to us

The machines are already talking.

You can either listen, or keep guessing. One option is free.