AIO score
Strong79 / 100
Top fixes
- High Reduce JS-only content for core pages (AI crawlers may not render JS) (Renderability & JS: 6/20)
- Low Add Subheadings (missing from the server HTML on 1 of 2 sampled pages) (Structure & readability: 23/30)
Score breakdown
- Technical & crawlability 25/25
- Renderability & JS (weakest) 6/20
- Structure & readability 23/30
- Metadata & consistency 25/25
What's working
- Technical & crawlability: 25/25
- Metadata & consistency: 25/25
- Structured data (JSON-LD) on all 2 sampled pages
- Every sampled page has an H1 heading
Ghost Test: can AI crawlers answer buyer questions?
We ask an AI model each question twice: once using only the HTML a crawler like GPTBot receives, and once using the page as visitors see it after JavaScript runs. A question is flagged when only the visitor version can answer it.
Homepage only: some of the pages the targeted questions need were not found, so we asked one broad question about the homepage instead.
What is this, and how does it make my life better? Gap
Page: https://abmagency.com/
Only the visitor version could answer: AI crawlers miss this information.
Answer from what the AI crawler sees
01 not found
Answer from what visitors see
01 The ABM Agency is a full-service agency specializing in account-based marketing (ABM) solutions, particularly 1:1 and 1:few campaigns. It helps businesses achieve transformative results by applying effective ABM processes, technologies, and data tailored to their unique needs. This can make your life better by streamlining your marketing efforts, enhancing engagement with high-potential accounts, and ultimately driving more revenue through personalized and targeted campaigns.
Page-by-page breakdown
Each sampled page, scored on what an AI crawler receives before any JavaScript runs.
- abmagency.com 65/100
homepage · 0% of visible text in server HTML
Content needs JavaScript Missing: Subheadings - abmagency.com/fr/ 75/100
product · 0% of visible text in server HTML
Content needs JavaScript
Audit scope
- Pages sampled
- 2
- How pages were found
- Homepage links only (no usable sitemap)
- Coverage · certainty
- partial · medium
Crawl warnings
- Couldn't load https://abmagency.com/pricing
- Couldn't load https://abmagency.com/integrations
- Couldn't load https://abmagency.com/trust
- Couldn't load https://abmagency.com/security
Compare to peers
- similarweb.com (no public report in this directory)
- zoominfo.com (audit report)
- leadiq.com (audit report)
- crunchbase.com (audit report)
- linkedin.com (audit report)
Can AI assistants open this site?
This report shows what AI crawlers read. It doesn’t test whether ChatGPT, Claude and Perplexity get past the site’s firewall when a person asks them about a page. If this is your company, the Threshold Test checks each assistant against your key pages and tells you what to change. It’s free once you’ve claimed this profile.
Report FAQ
What is the Ghost Test?
The Ghost Test checks whether an AI crawler can answer the questions buyers ask, such as pricing and main features. We ask an AI model each question twice about the page most likely to answer it: once using only the HTML the server sends (what crawlers like GPTBot read, since they don't run JavaScript), and once using the page as visitors see it. A question is flagged as a gap when only the visitor version can answer it.
Does a gap mean an AI hallucinated about this company?
No. A gap means the information exists on the site but AI crawlers can't read it without running JavaScript. That raises the risk that an AI assistant guesses or leaves the fact out, but it is not a record of a specific wrong answer.
How is the AIO score calculated?
The AIO score adds up four pillars: technical & crawlability (25 points), renderability (20), structure (30) and metadata (25). Each is scored on the HTML AI crawlers receive. The methodology page lists every check.
What do coverage and certainty mean?
Coverage describes how many of the site's important pages the audit managed to sample; certainty describes how consistent the evidence was. Lower values mean the report is directional and worth re-running.