Original Research · July 2026
Crawled July 28-29, 2026 · 526 sites attempted · 472 reachable · Kalamazoo, Portage & Battle Creek · aggregate data only
Key Findings
Customers increasingly ask ChatGPT and Google AI who to hire, and those systems read websites by machine. Of 472 Southwest Michigan business sites we audited, only 5 (1.1%) passed all three core AI-readiness checks: business schema, FAQ markup, and a sitemap. Just 3% use FAQ schema, the content format AI systems quote most, and half provide no machine-readable business identity at all. The region as a whole earns an AI-Readiness Index of 75/100, a C, and no city has an edge: Portage scores 76, Kalamazoo 75, Battle Creek 74. For local businesses, AI visibility is a wide-open competitive lane.
Share of audited sites passing each check. The basics are mostly covered. The AI-visibility layer collapses at the bottom of the list, and that is exactly the layer AI systems read.
A note on llms.txt: most of the 22% came from website builders that now generate the file automatically, not from a deliberate AI strategy. Only 4% of sites actively block AI crawlers, which means the door is open; almost nobody has walked through it.
Average AI-Readiness Index by industry. Dentists, law firms, and HVAC companies lead, likely a side effect of marketing-agency involvement in those verticals. Manufacturers and veterinarians, industries where word of mouth has always carried the load, trail the region.
Ranking on Google took years of content and links, and the winners built moats. AI answers are newer and still volatile. When 99% of your local competitors fail at least one critical machine-readability check, a few weeks of correct technical work puts you in the top handful of AI-visible businesses in your market.
These sites mostly look fine to a human. What is missing is the machine layer: schema that says who you are, where you are, and what you do, plus content structured as questions and answers. That is plumbing, not design, and it is fixable without rebuilding anything.
Website builders are already auto-generating llms.txt files, which is how 22% of sites have one without trying. Schema and FAQ automation will follow. The advantage belongs to businesses that move before readiness becomes a default setting.
On July 28-29, 2026, we ran an automated audit of 526 business websites headquartered in Kalamazoo, Portage, Battle Creek, and surrounding Southwest Michigan communities, compiled from public directories and local listings across roughly 19 industries, with 140+ sites per core city (Kalamazoo, Portage, Battle Creek). 472 sites were reachable. National chains, franchises, and our own clients were excluded.
Each site received four polite requests (homepage, robots.txt, llms.txt, sitemap.xml) from an identified research crawler. We checked: HTTPS, mobile viewport markup, robots.txt presence and AI crawler rules, meta descriptions, title tag quality (20 to 75 characters), XML sitemap validity, schema markup types, FAQPage markup, llms.txt (each positive re-verified against a control path to eliminate soft-404 false positives), and server response time (median: 0.50s).
Results are published in aggregate only; no individual business is named or scored publicly. The same checks power our free AI Visibility Grader, so any business can run its own site against this benchmark. Media and researchers: we are happy to share additional aggregate breakdowns on request at contact@warriorwebco.com.
AI-search readiness means a website provides the signals AI systems like ChatGPT, Google AI Overviews, Gemini, and Perplexity use when deciding which businesses to name in an answer: structured data (schema markup) identifying the business, question-and-answer content in FAQ markup, a sitemap for discovery, and open crawler access. A site can look great to humans and still be unreadable to these systems.
On July 28 and 29, 2026 we ran an automated audit of 526 Southwest Michigan business websites across roughly 19 industries in Kalamazoo, Portage, Battle Creek, and surrounding communities, drawn from public directories and local listings. 472 were reachable. Each site was checked for HTTPS, mobile markup, robots.txt and AI-crawler access, meta descriptions, XML sitemaps, schema markup types, FAQ markup, llms.txt (re-verified manually against false positives), title quality, headline structure, and server response time, then scored 0 to 100 using the same weighted methodology as our public AI Visibility Grader. Results are reported only in aggregate; no individual business is named.
The AI-Readiness Index is the average weighted website score for a market, on a 0 to 100 scale, computed with the same methodology as our free AI Visibility Grader: structured data and FAQ markup carry the most weight, followed by crawlability, titles, mobile readiness, and speed. In July 2026, Southwest Michigan as a region scores 75, a C: Portage 76, Kalamazoo 75, Battle Creek 74.
AI systems quote question-and-answer content more than any other format, and FAQ schema is how a website labels that content so machines can extract it reliably. Marked-up pages get cited in AI answers at substantially higher rates. In our audit only 3 percent of Southwest Michigan business sites used it, which makes it one of the cheapest competitive advantages available in this market right now.
Run it through our free AI Visibility Grader at warriorwebco.com/tools/ai-visibility-grader. It performs the same checks and the same scoring as this study on any site you enter, instantly, with no email required, and explains how to fix anything it flags.
Run your site through the same checks this study used, free and instant, no email required. If you want the gaps fixed, we are the first agency in this market doing AI search optimization, and we proved the playbook on our own product first.