The State of CPA Firm Websites 2026: Structured-Data and AI-Legibility Dataset (n=556)

This dataset accompanies the study “The State of CPA Firm Websites 2026” by Axion Deep Digital. It measures the structured-data legibility of United States accounting-firm websites: whether their sites make named experts, identity, and answer-shaped content machine-readable for search engines and AI answer engines. Sample: 556 unique accounting-firm domains sampled from OpenStreetMap (office=accountant or office=tax_advisor with a website tag) across 68 U.S. metros, stratified into major, tech-hub, mid-size, and small markets. Each firm’s homepage, robots.txt, and llms.txt were fetched in a single pass and parsed for machine-readable signals. Adoption statistics are computed over the 464 reachable firms, defined as sites returning HTTP 200 with a non-empty body. Headline findings among reachable firms, using Wilson score 95% confidence intervals: • 94.2% publish no Person schema (95% CI: 91.7–96.0).• 62.1% have no identity schema.• llms.txt is present on 19.8% of sites, but only 9.3% is firm-authored. The remainder is auto-generated by Wix, Shopify, or SEO plugins.• 2.4% block at least one AI crawler. The study measures structured-data legibility, not firm quality, and makes no causal claims. The data is anonymized at the row level. Domain and city information have been removed so that no individual firm is identifiable. Columns: site_id: Anonymized identifier. stratum: Market classification: major, tech, mid, or small. platform: Detected CMS or website builder. reachable: 1 if the site returned HTTP 200 with a non-empty body. All adoption statistics use reachable=1. home_status: Homepage HTTP status code. has_identity_schema: 1/0 JSON-LD flag indicating identity-related schema. has_trust_schema: 1/0 JSON-LD flag indicating trust-related schema. has_qa_schema: 1/0 JSON-LD flag indicating question-and-answer schema. has_person_schema: 1/0 JSON-LD flag indicating Person schema. llms_present: 1 if a valid /llms.txt file is present. llms_provenance: Source classification: manual, Wix, Shopify, AIOSEO, Yoast, RankMath, or generated_other. llms_platform_generated: 1 if the llms.txt file was generated by a platform or plugin rather than authored by the firm. ai_bots_blocked_count: Number of blocked crawlers among the 14 tracked AI crawlers. schema_types: Semicolon-separated JSON-LD @type values. Full report, methodology, and charts: https://www.axiondeepdigital.com/research/state-of-cpa-websites-2026