Identity needed clearer structure
565 assessable homepages lacked the specific Organization schema signal checked in this study.
The 2026 AI Website Readiness Study
Most local business websites work for people. They still make machines guess.
We examined 1,000 public homepages across 10 industries and 10 major U.S. metros. Most supplied useful evidence—but important identity, structure, and representation signals were often incomplete.
1,000websites submitted
897assessable homepages
61median readiness score
10U.S. metro areas
Executive summary
Functional websites still leave machines to infer too much
The clearest pattern was not that these sites were universally “bad.” Most were functioning business websites. The recurring gap was that important identity, metadata, and representation signals were incomplete or inconsistent for systems expected to retrieve and summarize them.
A middle-range score does not mean a business is invisible. It means the captured homepage supplied useful evidence while leaving concrete opportunities to reduce ambiguity.
The short version
Three numbers explain the opportunity
Shared previews were incomplete
221 assessable homepages had incomplete Open Graph metadata for social and messaging previews.
Broad crawler blocks were uncommon
Only 31 assessable sites blocked at least one major retrieval crawler tracked by BoostGeo.
The takeaway: the larger opportunity is usually clearer, consistent evidence—not simply opening the door to more crawlers.
The findings
Read the evidence, not a grade
Every chart states its denominator. Unable-to-assess sites remain visible in coverage reporting and are excluded from score and finding rates.
Score distribution
Most assessable sites landed in the middle
280–1928 sites
15320–39153 sites
23940–59239 sites
36060–79360 sites
11780–100117 sites
Common opportunities
Explicit organization identity was the largest recurring substantive gap
Organization schema opportunity565 of 897
63.0%WebSite schema opportunity336 of 897
37.5%Incomplete Open Graph metadata221 of 897
24.6%No JSON-LD detected180 of 897
20.1%No meta description detected117 of 897
13.0%No H1 detected103 of 897
11.5%At least one major crawler blocked31 of 897
3.5%Evidence coverage
Unable results were separated—not scored as zero
High-confidence evidence1,000 submitted
766 high confidence 76.6%
All assessable evidenceincludes 131 limited-confidence results
897 assessable 103 unable
Industry context
Descriptive medians varied across the ten sampled industries
Accounting and CPA firmsmedian Website Readiness
48General dentistrymedian Website Readiness
65HVAC contractorsmedian Website Readiness
63Insurance agenciesmedian Website Readiness
60.5Med spasmedian Website Readiness
66Personal-injury lawmedian Website Readiness
69Plumbing contractorsmedian Website Readiness
56Residential real estatemedian Website Readiness
52Roofing contractorsmedian Website Readiness
61Veterinary clinicsmedian Website Readiness
58How to read the report
Definitions behind the numbers
- Website Readiness
- How clearly the captured homepage exposed retrievable, interpretable evidence. It is not an AI mention or traffic score.
- Assessable
- BoostGeo established enough representative homepage evidence to calculate a score.
- Opportunity
- A rule detected a specific signal that could be clearer or more complete. It does not automatically mean the site is broken.
- Unable
- The collection did not substantiate representative evidence. Unable results were kept separate—not converted to zero.
Identity was the most common substantive opportunity
BoostGeo found a medium-priority Organization schema opportunity on 565 of 897 assessable homepages (63.0%). This does not mean those businesses were unidentifiable. It means the captured homepage lacked the specific Organization JSON-LD signal checked in this study.
Preview and website metadata were frequently incomplete
Three hundred thirty-six sites had a WebSite schema opportunity, while 221 had incomplete Open Graph metadata. These signals do not guarantee rankings or AI citations. They give retrieval and preview systems cleaner, more consistent evidence.
Foundational omissions affected a meaningful minority
One hundred eighty assessable homepages had no detected JSON-LD, 117 had no detected meta description, and 103 had no detected H1. These are homepage observations—not declarations about every page on each domain.
Crawler blocks were uncommon, but worth reviewing
Thirty-one assessable sites blocked at least one major retrieval crawler tracked by BoostGeo. A block may be intentional and does not imply that all AI access is blocked. Policies should match the organization’s publishing goals.
Continue with the evidence
Turn the findings into a useful next question
The human story
Can AI understand your business website?
What the study reveals in plain English—and what Website Readiness cannot prove.
Read the story →Industry findingsHow the ten industry samples compared
Explore descriptive medians without turning them into misleading business rankings.
Explore industries →Practical guideMake your website easier for AI to understand
A plain-English order of work based on the recurring evidence gaps.
Use the checklist →What to do first
Improve clarity before chasing technical perfection
- Confirm the basics.State who you are, what you offer, where you operate, and what visitors should do next.
- Align the page signals.Make the title, description, H1, canonical URL, and social preview tell the same accurate story.
- Add structured identity.Use accurate Organization and WebSite markup that agrees with visible content.
- Review crawler policy.Allow or block retrieval intentionally—based on publishing goals, not fear or default settings.
Check your website freeSee evidence first. No technical setup required.
Methodology
A frozen cohort and an evidence-gated score
The frozen cohort included 100 websites in each of 10 industries and 100 in each of New York, Los Angeles, Chicago, Dallas–Fort Worth, Houston, Miami, Atlanta, Phoenix, Seattle, and Boston. Directories, aggregators, duplicate domains or brands, social-only profiles, and founder-controlled test sites were excluded before scoring.
Sites were evaluated with BoostGeo Website Readiness scorer SIGIL/λ-11 on August 27–28, 2026. The scorer withholds a numeric score when it cannot establish representative page evidence. The primary analysis contains 766 high-confidence and 131 limited-confidence results; 103 unable results remain separate. The high-confidence-only median was 60 versus 61 across all assessable sites, with the same upper quartile of 72.
Sample registerSee the 1,000-domain cohortPlain-text CSV · no rankingsChart dataReuse every published totalDistribution, findings, coverage, and industry medians
Study limitations
- The quota-based cohort is descriptive, not statistically representative of every local business or U.S. website.
- The study evaluated captured homepage evidence, not every page on each domain.
- Website Readiness is not independent public AI visibility.
- Industry medians are context, not league tables or causal comparisons.
- Public websites can change after collection.
- No traffic, lead, revenue, conversion, or customer-interview data was used.