Open Dataset: AI Crawler & Schema Signals, 100 B2B SaaS (2026)

Evidence disclosure

Original Research

Observation date

Aug 24, 2026

Method

Deterministic public HTTP inspection of 100 B2B SaaS domains for AI crawler access, llms.txt presence, canonical and sitemap hygiene, and four schema.org types. No language model was used to score or rank companies.

Scope

The complete frozen 100-company sample behind Growthract's 2026 AI Search Readiness Benchmark, collected August 24, 2026.

What this supports

Independent reanalysis of the same deterministic technical signals for the same 100-company sample.
Citation of specific figures back to a single, stable source page.

Evidence boundary

That any figure derived from this dataset implies causation for AI citation, ranking or recommendation outcomes.
That this sample represents the entire B2B SaaS market.
That the underlying signals have not changed since the August 24, 2026 collection date.

Domain-level CSV, sampled page-level CSV and a JSON summary are downloadable directly from this page and the main benchmark report.

Full benchmark report

Direct answer

Where can I download the raw data behind Growthract's 2026 AI Search Readiness Benchmark?

The full domain-level CSV, sampled page-level CSV, and a machine-readable JSON summary from Growthract's 2026 AI Search Readiness Benchmark (100 B2B SaaS companies, AI crawler access, llms.txt, and four schema types) are published openly and linked directly from this page for reuse, citation and independent analysis.

This is the canonical entry point to the raw dataset — direct file links, a field dictionary, and a suggested citation format.

The data is deterministic, publicly observable technical evidence, not model-scored or AI-generated.

Reuse and cite freely; we ask that you link back to the source benchmark rather than republishing the raw numbers uncredited.

This page exists for one reason: to be the direct, stable link to the raw data — not another write-up of the findings.

If you're citing a specific figure, reanalyzing the sample, or building your own visualization, start here.

Download the data

domains-final.csv ↓

Domain-level evidence for all 100 companies: AI crawler access per bot, llms.txt state, schema presence, canonical and sitemap hygiene.

pages-final.csv ↓

Page-level evidence from the deterministic supporting-page sample beyond just the homepage.

benchmark-summary.json ↓

Machine-readable aggregate summary, including the four-category stratum breakdown.

Field dictionary (domains-final.csv)

company / stratum

Company name and its assigned category out of four stratified groups of 25.

robots_state / robots_http_status

Whether robots.txt was fetched successfully, and its HTTP status.

{bot}_status

One column per tested crawler (oai_searchbot, gptbot, claudebot, perplexitybot, google_extended, and others) — allowed, blocked or unknown.

llms_txt_state / llms_txt_present

Whether /llms.txt exists, and whether it follows a structured format.

homepage_jsonld_present

Whether any application/ld+json markup was found in the initial homepage HTML.

homepage_organization_schema / _website_schema / _software_application_schema

Presence of each specific schema.org type on the homepage.

canonical_present / canonical_self_referencing

Whether a canonical tag exists, and whether it points to the page itself.

sitemap_present / sitemap_url_count

Whether a sitemap was found, and how many URLs it listed.

Full column definitions and collection notes are in the accompanying README published alongside the CSVs.

How to cite this dataset

Use freely, and please link back to the source benchmark rather than republishing figures without attribution:

Growthract. (2026). 2026 AI Search Readiness Benchmark: 100 B2B SaaS Websites [Dataset].
https://www.growthract.com/insights/blogs/ai-search-readiness-benchmark-2026

Cuts of this data we've already published

Before you rebuild an analysis from scratch, check whether we've already published it:

Limitations to carry into any reuse

This is a frozen, point-in-time sample of 100 companies, not a census of the B2B SaaS market. Unknown values were excluded from the relevant denominator rather than counted as failures — treat any percentage you compute the same way. Full methodology and limitations are documented in the main benchmark report.

Beyond the aggregate sample

Get the same signals checked for your own site.

Run Growthract's free diagnostic to check the same crawler access, llms.txt and schema signals against your own website.

Run your free audit Read the full benchmark

FAQs

Can I republish charts built from this data?

Yes — please link back to this page or the main benchmark report as the source.

Will this dataset be updated?

It's currently a frozen snapshot from August 24, 2026. Any future refresh will be published as a dated update rather than silently overwriting these files.

What format are the files in?

Two are standard CSV files, and the summary is a machine-readable JSON file — all directly downloadable, no account or request required.

Continue exploring

Also relevantExplore Growthract’s SEO + AEO approach

Case study

A public technical audit of growthract.com covering 49 production HTML pages, crawler access, robots.txt, sitemap integrity and legacy routes — with the limits of the evidence stated explicitly.

See the case study →