Case Study: How a Windows & Doors Company Got Cited by AI in 5 Markets

Here's a question most home-services owners can't answer: when someone asks ChatGPT "who does the best window replacement near me," does the AI ever say your name? For one windows & doors company we work with, the honest answer used to be no — not in their home market, and not in any of the surrounding towns they served. The AI either named a competitor or named nobody at all. Over a stretch of months, that changed: the company went from never being named by AI to being cited across roughly five of its local markets, with the majority of the major AI engines naming the brand when asked about it directly. This is the story of what we changed and why it worked — told with the numbers we can stand behind and nothing we can't. (This case study is anonymized and shared as a draft pending the client's sign-off, in keeping with how we handle every client's privacy.)

What happened in this case study?

A windows & doors company went from never being named by AI to being cited across about five of its local markets over a period of months — with the majority of the major AI engines naming the brand when asked about it directly. We're sharing this as anonymized, directional proof, not a guarantee. Different businesses, markets, and starting points produce different results.

Let's be precise about what those words mean, because vague case studies are exactly what we'd warn you about. "Cited by AI" here means the brand showed up by name when an AI engine answered a question about it — not that it ranked #1 on Google, and not that we can promise the same outcome for you. "Five local markets" means the company's home town plus the surrounding service areas where it does business.

And one honest caveat up front: AI answers move. An engine might name a business one week and phrase its answer differently the next. So when we say the brand is now named by most major engines, we mean it held up across repeated checks over time — not a single lucky screenshot. That distinction is the whole reason we measure instead of guess.

The problem: a good company, invisible to AI

The company did solid work and had a real website — but when customers asked an AI for a recommendation, the brand simply didn't come up. Their pages weren't written or structured in a way an AI could quote, and the wider web didn't tell the engines clearly enough who they were, where they worked, or what they did.

This is more common than it sounds. By one industry measure, ChatGPT recommends only about 1.2% of local businesses when asked — meaning the overwhelming majority of good, established companies are invisible in AI answers entirely (SOCi 2026 Local Visibility Index). Being good at your trade and being named by an AI are two different things, and the gap between them is exactly where this company was stuck.

The root causes weren't exotic. Their service pages buried the answer to "what do you do and where" under marketing language. The site didn't directly answer the plain questions customers ask before a window replacement. And outside the site — the reviews, listings, and mentions an AI leans on to decide a business is real and trusted — the signals were thin and inconsistent across markets. To an AI trying to write a confident answer, there wasn't enough to grab onto.

What we actually changed

We did the unglamorous work in order: rewrote the pages to answer real customer questions clearly and high on the page, structured each page so a machine could quote it, and strengthened the company's credibility across the web — accurate listings, reviews, and consistent business details in every market they serve. No tricks, no black boxes.

The same three steps run through every project we take on:

Step 1

Find out what's happening

Check where the brand stands on Google and in AI today, market by market. Nothing assumed.

Step 2

Build what's missing

Rewrite and structure pages to answer real questions — built for Google and AI at once.

Step 3

Prove it's working

Check the AI engines over time and report real numbers — what moved, what didn't.

On the pages themselves, the biggest shift was simple: lead with the answer. A customer searching for window replacement wants to know what you install, where you work, and roughly what to expect — so we put that plainly near the top of each page instead of three paragraphs down. AI engines quote what they can lift cleanly, and burying the answer hides it from both the reader and the machine. If you want the full breakdown of this approach, our explainer on what AI SEO actually is walks through the mechanics.

Just as important was the work off the site. AI engines decide whether a business is real and trustworthy partly by what the rest of the web says about it — reviews, accurate listings, consistent name-address-phone details in every town it serves. We made sure those signals lined up across all five markets, so the engines had a clear, consistent picture to draw from instead of a muddle.

The results, told honestly

Over a period of months, the company went from not being named by AI at all to being cited across about five of its local markets, with most of the major AI engines naming the brand for questions about it. We're sharing these as directional figures from one engagement — real, but not a promise that your numbers will match.

Here's the part where a lot of agencies would inflate things, so we'll be plain about what these results are and aren't:

• What changed: from never named by AI, to named across roughly five local markets — a genuine shift from invisible to present in the answer.
• The timeframe: this happened over a period of months, not overnight. AI visibility builds; it doesn't flip like a switch.
• What we're not claiming: we're not putting a revenue figure or a "300% more leads" headline on this. Those numbers are easy to fake and hard to verify, so we don't lead with them. We can stand behind the visibility shift because we checked it ourselves, repeatedly, over time.

That last point matters. There's a meaningful difference between an AI using your content as a source and an AI naming your business in its answer — and across the industry, getting named by name is the rarer, more valuable outcome. Semrush's 2025 AI Visibility Index calls this the "mention-source divide" — fewer than one in five brands manage both being cited as a source and being named in the answer (Semrush, 2025). Moving a brand into the "named" column across multiple markets is the result worth talking about, and it's the one we measured.

Why it worked — and what didn't move fast

It worked because the work matched how AI engines actually decide who to name: clear answers, clean structure, and real credibility across the web. There was no shortcut and no secret tactic. What didn't move fast was exactly what you'd expect — visibility built over months, market by market, not in the first few weeks.

The honest version of "why it worked" is unremarkable, and that's the point. AI engines reward the same things a good neighbor would: be easy to understand, answer the question that was actually asked, and be the kind of business other people vouch for. We didn't game anything. We made a good company legible to the machines deciding which businesses to recommend.

And the parts that lagged are worth naming too. Some markets came online before others. AI answers shifted week to week before they settled into naming the brand consistently. And no single change carried the day — it was the page rewrites, the structure, and the off-site credibility working together. Anyone promising you instant AI visibility from one trick is selling you the buzzword, not the work. The real thing takes the right steps in the right order, and a bit of patience while it compounds.

How to check your own business — and what it'd take

Start by asking the AI engines about your own business the way a customer would: "best [your trade] in [your town]," or "who does [your service] near me." If your name doesn't come up — or a competitor's does — that's the gap, and it's the same gap this company closed. The fix is the same order of work: clear pages, clean structure, real credibility.

You can do a rough version of this yourself in ten minutes. Open ChatGPT, Perplexity, and Google's AI answers, ask each one a question a customer would ask in your market, and note whether you're named, a competitor's named, or nobody is. Do it for each town you serve. That alone tells you where you stand — and most owners are surprised by the answer.

If you'd rather have it checked properly and handed to you in plain language, that's exactly what our free Website Scorecard does — a snapshot of whether AI names you or your competitor for the searches that matter in your market, no sign-up wall. And if you want to understand the tools behind this kind of tracking, our rundown of the best AI visibility tools covers how the measurement works under the hood. Visibility is not luck. It is a system — and this case study is one example of that system doing its job.