We Audited 20 Small Business Websites
Here's What We Found
We didn’t run a survey. We didn’t ask anyone how they felt about their website. We searched — organically, across four verticals — and evaluated 20 real small business websites against the signals AI systems actually use to decide who gets recommended. No fluff, no hypotheticals. Here’s what we found.
Why We Did This — And How
AI-powered search is no longer a future problem. Google’s AI Overviews, ChatGPT, Perplexity, and Bing Copilot are actively replacing the first click for millions of queries every day. When someone asks “who’s the best mortgage broker near me” or “top youth soccer leagues in [city],” an AI is making a recommendation before any human ever sees a search result page.
We selected five websites from each of four verticals: B2B professional services, real estate and mortgage, restaurants and hospitality, and athletic leagues and organizations. We evaluated each one against the core signals AI systems use — structured data, named authorship, outcome-based content, external citations, and what we call trust signal architecture.
The results were consistent enough to be uncomfortable. Across all four verticals, the pattern was the same: these websites were built for people who already found them — not for an AI deciding who to surface first. A proper AI visibility audit exposes exactly that gap, and we saw it in nearly every site we reviewed.
Finding #1: B2B Consulting Sites Are Built to Impress, Not to Answer
Of the five B2B professional services firms we reviewed, zero had meaningful schema markup. One had a basic Organization schema that hadn’t been updated since the site launched. None had named authors on service pages or thought leadership content. Most had bios buried three clicks deep — if at all.
The content architecture was almost universally the same: a homepage that described the firm’s capabilities, a services page that listed offerings, and a contact form. Beautiful, professional, and completely opaque to an AI trying to match a client query to an expert.
When a business owner types “how do I reduce churn in a SaaS business” into an AI tool, that AI is looking for a named expert, a body of work, and outcome-based evidence. It’s not looking for “We offer strategic growth consulting for mid-market technology companies.” That sentence tells an AI nothing actionable.
The one exception in our sample was a firm that had reorganized its entire content strategy around client problems rather than service categories. Their blog answered specific questions their clients were actually asking. Their consultants were named and linked to the content they wrote. That firm appeared in AI-generated results. The others did not.
This is the E-E-A-T gap — Experience, Expertise, Authoritativeness, Trustworthiness — and heading into E-E-A-T 2026, it’s only getting more consequential. AI systems are trained to weight these signals heavily. If your site doesn’t surface them, you don’t exist in that decision layer. Learn more about how AI visibility works and why it’s reshaping who gets found first.
Finding #2: Mortgage Brokers Have a Great Story — Locked in a Format AI Can’t Read
Independent mortgage brokers have a genuinely compelling value proposition. They can shop across multiple lenders, often secure lower rates than captive bank loan officers, and provide a level of personalized service the big institutions simply don’t offer. It’s a strong argument — and almost none of them are making it in a way AI can extract and repeat.
Every mortgage broker website SEO review in our sample turned up the same issues. The differentiators were buried in paragraph text with no structured data to signal their importance. Loan officers were listed by name on a team page but not connected to any authored content, testimonials, or credential citations. None of the sites had LoanOfficer or Person schema. None cited professional associations or press coverage inline — even when those citations existed elsewhere on the web.
When an AI is asked “should I use a mortgage broker or go directly to a bank,” it’s going to pull from sources that have structured, citable answers. If your website reads like a brochure — even a good brochure — you’re not in that pull. You’re invisible at the exact moment a prospect is forming their decision.
The fix isn’t complicated, but it requires intent. Named loan officers need to be authors, not just headshots. The broker’s value proposition needs to live in structured, quotable content. Schema needs to reflect credentials. And external validation — lender relationships, association memberships, press mentions — needs to be surfaced on the site, not just assumed.
Finding #3: Restaurant Sites Are Optimized for the Wrong Moment
Restaurant websites are often genuinely impressive — beautiful photography, well-designed layouts, evocative copy. They’re built for the moment someone has already decided they want to eat there and is looking for the address. They are not built for the moment an AI is deciding which restaurant to recommend in the first place.
Not one of the five restaurant sites we reviewed had restaurant schema markup implemented correctly. Three had no structured data at all. Two had PDF menus — which are completely invisible to AI crawlers and search engines alike. None had a chef or owner story in a format AI could parse and attribute. None had cuisine, price range, or reservation schema that would allow an AI to confidently include them in a response to “best Italian restaurants in [city] for a date night.”
The restaurants that DO appear in AI recommendations? In nearly every case, they got there through external media coverage — a local food blog review, a magazine feature, a well-structured Google Business Profile — not through anything on their own website. Their site was a passenger. The media was doing the heavy lifting.
That’s a fragile position. Media coverage fades. Profiles go stale. The restaurant’s own website — the one asset they fully control — should be the most authoritative source about what they offer, who they serve, and why they’re worth the trip. Right now, for most independent restaurants, it’s the weakest link in their digital presence.
Finding #4: Youth Sports Leagues Are Invisible by Design
This one hurt to document. Youth sports league website design is almost entirely built for existing members — parents who already registered, coaches who need the schedule, board members who know the login. For a new family moving into the area and asking an AI “youth soccer leagues near me for a 9-year-old,” these sites offer almost nothing.
Game schedules were embedded via Google Sheets in three of the five sites we reviewed. AI cannot read embedded spreadsheets. There was no SportsOrganization schema, no Event schema for games or tryouts, no named leadership, and no content explaining what the league is, who it’s for, or what makes it a good fit for a family. The sites assumed you already knew all of that.
The practical consequence is registration loss. National franchise leagues — your Regals, your i9 Sports, your US Youth Soccer affiliates — have professional digital infrastructure with proper schema, keyword-optimized content, and structured data that makes them easy for AI to recommend. An independent league with 20 years of community history is losing to a franchise that opened last year, simply because the franchise built its site for discovery and the independent league built its site for administration.
Named coaches and directors, structured event data, content that explains the league’s philosophy and age divisions, and basic SportsOrganization schema would move the needle significantly. It’s not a technology problem. It’s a strategy problem.
The Throughline Across All Four Verticals
Every website we reviewed was built with a specific user in mind: someone who had already found the business, already decided they were interested, and needed confirmation or contact information. That’s a reasonable website to build — in 2015.
In 2026, there’s a new gatekeeper between your business and that prospective customer. It’s an AI system that processes a query, evaluates available sources, and makes a recommendation before any human interaction occurs. That AI has no patience for beautiful photography without alt text, PDF menus it can’t parse, or paragraph copy that describes your values without ever stating your outcomes.
The businesses that will win AI-era search are the ones that make it effortless for an AI to understand: who you serve, what problem you solve, what makes you credible, and what someone should do next. That requires structured data, named expertise, outcome-focused content, and external validation signals — all working together. A thorough AI visibility audit for your small business website AI search presence is the starting point. You can’t fix what you haven’t mapped.
Heading into E-E-A-T 2026, the gap between optimized and unoptimized sites will not close on its own. It will widen — because the businesses that understand this are moving fast, and the ones that don’t are still updating their PDF menus.
FAQ: What Business Owners, Parents, and Restaurateurs Are Actually Asking
How do I know if my small business website shows up in AI search results?
The most direct method is to search for the queries your customers actually use in ChatGPT, Perplexity, and Google’s AI Overview and see whether your business appears. If you don’t show up — or a competitor consistently does — that’s your answer. A structured AI visibility audit will map the specific signals you’re missing: schema, authorship, citation gaps, and content structure. Most small businesses find they have the raw ingredients but haven’t organized them in a way AI systems can use.
Why isn’t my restaurant showing up when people search for restaurants near me on AI tools?
Almost always, it’s a combination of missing restaurant schema markup, an unstructured menu format (PDF menus are invisible to AI), and the absence of a chef or owner narrative that an AI can attribute and cite. AI tools pull from sources that are easy to parse and verify. If your website doesn’t have structured data communicating your cuisine type, price range, hours, and location in machine-readable format, you’re relying entirely on external sources — reviews, press, your Google Business Profile — to carry you. That works until it doesn’t.
How can my youth sports league compete with national franchise programs online?
The national programs win on infrastructure, not on quality. They have SportsOrganization and Event schema, keyword-aligned content about age divisions and registration, and named leadership with structured bios. Independent leagues can match all of that — the technology is the same. What’s required is intentional restructuring of your website away from member administration toward public discovery. That means readable schedules (not embedded spreadsheets), content that explains your league’s philosophy and who it’s for, and structured data that tells AI systems exactly what you offer and when registration opens.
Ready to See Where You Stand?
If anything in this post sounded familiar, your website has the same gap we found in all 20 sites we reviewed. The good news: it’s fixable, and the businesses that fix it now have a real window before their market catches up.
We run structured AI visibility audits for small businesses across all four verticals covered here — and we build the fix, not just the report. Book a discovery call and we’ll show you exactly where you’re invisible and what it would take to change that.
BOOK YOUR DISCOVERY CALL