Every number on this page is dated, names the exact keyword and its monthly search volume, and identifies the page that earned the position. You can run the same searches yourself and check them. That is the standard we hold our reporting to for clients, so it is the standard this page is held to as well.
Clients are described by what they do rather than by name, at their request and ours. A business that has paid for search work does not always want its competitors reading the playbook.
These AI SEO case studies record two client engagements run through the Rank Intelligence Engine, with positions measured daily in Google and reported against a baseline taken before the work started. A bathtub reglazing company in Los Angeles moved four tracked commercial keywords from an average position of 5 in February 2026 to an average of 1.75 by 27 August 2026, holding thirteen keywords in the top five. A mobile outboard motor repair shop in Fort Lauderdale moved six tracked keywords from an average position of 8 on 1 August 2026 to an average of 2.5 twenty-six days later, holding eighteen keywords in the top five.
Case Study One: A Bathtub Reglazing Company in Los Angeles
The situation
A refinishing contractor competing in one of the most contested local markets in the country, against national franchise operators with far larger budgets and far older domains. Reglazing is a considered purchase with a narrow window: a homeowner decides to refinish rather than replace, searches once, and calls someone within a day or two. Being on page one is not enough in that window. Being in the first three results is the whole job.
What the Rank Intelligence Engine found
The commercial head terms in this niche cluster tightly around a handful of phrasings that mean the same thing to a customer and different things to Google: reglazing, refinishing, resurfacing, and every ordering of tub, tile and bathtub around them. A site that spreads those across several competing pages splits its own signal. A site that concentrates them on one authoritative page and gives genuinely separate services their own pages does not. The scan also showed adjacent demand the site was not addressing at all, in countertop work and in accessibility conversions.
What was built
The homepage was made the single owner of the reglazing head family, with the competing variants consolidated into it rather than fighting it. Countertop refinishing was given its own page to own its own family. A page was built for senior access conversions, a service the business performs and had never published. Each page was structured so that a specific question has a specific, quotable answer on it, which is what earns a position in Google and a citation in an AI answer.
The numbers, measured in Google Los Angeles
Four tracked commercial keywords, same keyword records on both dates. Average position 5 on 1 February 2026. Average position 1.75 on 27 August 2026.
| Keyword | Monthly searches | 1 February 2026 | 27 August 2026 | Page that ranks |
|---|---|---|---|---|
| reglazing tub and tile | 590 | 5 (set average) | 1 | homepage |
| tile reglazing company | 50 | 5 (set average) | 1 | homepage |
| tub tile reglazing | 480 | 5 (set average) | 2 | homepage |
| reglazing bathroom tile near me | 20 | 5 (set average) | 3 | homepage |
The wider picture on the same day: thirteen keywords in the top five, spread deliberately across three pages rather than piled onto one.
| Keyword | Monthly searches | 27 August 2026 | Page that ranks |
|---|---|---|---|
| reglaze tub and tile | 590 | 1 | homepage |
| tub glazer | 880 | 4 | homepage |
| bathtub and tile refinishing | 480 | 4 | homepage |
| refinish tub and tile | 480 | 5 | homepage |
| reglazing countertops | 140 | 4 | countertop refinishing page |
| counter reglazing | 140 | 5 | countertop refinishing page |
| reglaze tile countertops | 70 | 4 | countertop refinishing page |
| senior access conversions | new term | 1 | senior access page |
What this case demonstrates
Consolidation beats proliferation. The instinct of most agencies facing a cluster of similar keywords is to build a page for each one, which produces a set of pages that compete with each other and none that wins. The opposite move, concentrating a family on one page and giving genuinely different services their own, is what moved these positions. It is also what makes a page citable: an AI engine looking for the best answer about tub and tile reglazing finds one authoritative page rather than six thin ones.
Case Study Two: A Mobile Outboard Motor Repair Shop in Fort Lauderdale
The situation
A mobile marine mechanic in the largest recreational boating market in the United States, competing against established dealerships and marina service departments. The customer is a boat owner whose engine has failed, often that morning, searching on a phone from a dock. Speed of decision is extreme and the shortlist is short.
What the Rank Intelligence Engine found
Marine repair demand splits along three axes at once: the service performed, the engine brand, and whether the mechanic comes to the boat. Those are not variations of one search. A boat owner typing outboard mechanic near me and one typing evinrude outboard service want different things, and a single page trying to answer both answers neither convincingly. The scan mapped which axis carried real volume and which pages would need to exist to cover them without overlapping.
What was built
Four purpose-built pages, each owning one field. A Fort Lauderdale boat repair page for the geographic and general repair demand. A mobile marine mechanic page for the come-to-the-boat searches. Brand pages for the two engine makes with the strongest demand. Each was engineered rather than templated: the questions a boat owner actually asks, answered directly enough to be lifted.
The numbers, measured in Google Fort Lauderdale
Six tracked keywords, same keyword records on both dates. Average position 8 on 1 August 2026. Average position 2.5 on 27 August 2026, twenty-six days later.
| Keyword | Monthly searches | 1 August 2026 | 27 August 2026 | Page that ranks |
|---|---|---|---|---|
| outboard mechanic | 590 | 8 (set average) | 3 | Fort Lauderdale boat repair page |
| mobile outboard mechanic | 140 | 8 (set average) | 2 | mobile marine mechanic page |
| outboard engine repair | 100 | 8 (set average) | 2 | Fort Lauderdale boat repair page |
| small outboard motor repair near me | 210 | 8 (set average) | 2 | Fort Lauderdale boat repair page |
| outboard boat repair near me | 40 | 8 (set average) | 3 | Fort Lauderdale boat repair page |
| outboard mechanic near me | 20 | 8 (set average) | 3 | Fort Lauderdale boat repair page |
Movement was still running on the day these numbers were taken. Four keywords changed position in the preceding twenty-four hours alone.
| Keyword | Monthly searches | 27 August 2026 | Change in one day |
|---|---|---|---|
| outboard motor repair near me | 2,400 | 4 | up 7 places in a day |
| outboard motor repairs near me | 290 | 3 | up 7 places in a day |
| yamaha outboard mechanic near me | 210 | 4 | up 3 places in a day |
| evinrude boat motor repair | 480 | 5 | up 2 places in a day |
Eighteen keywords sat in the top five that day. The distribution is the part worth studying, because it is the mechanism rather than the outcome.
| Page | Keywords it holds in the top five |
|---|---|
| Fort Lauderdale boat repair page | 10 |
| Evinrude outboard repair page | 4 |
| homepage | 2 |
| mobile marine mechanic page | 1 |
| Suzuki outboard repair page | 1 |
What this case demonstrates
Rankings were engineered into purpose-built pages, not bolted onto a homepage. One page carries ten of the eighteen positions because it was built to carry exactly that field. The brand pages carry their own. Nothing is competing with anything else. This is the architecture that both Google and the AI engines reward, and it is the reason the movement happened in twenty-six days rather than six months.
What Both Cases Have in Common
Neither engagement bought a link. Neither published a page for a city the business does not serve. Neither claimed a review rating that did not exist. In both, the work was the same four stages: the standing research into what the algorithm and the AI engines currently reward, a scan of the site and its competitors against that research, a sequenced map of what to build and in what order, then implementation with monthly measurement.
And in both, the decisive move was architectural rather than cosmetic. Deciding which page owns which field, and then building that page properly, produced more movement than any amount of tweaking would have. That is the finding we would most want a prospective client to take from this page.
How to Read These Numbers
Positions move daily. Every figure here carries the date it was recorded. Check them yourself and you may see something slightly different, because that is how search works. Any agency presenting rankings without dates is presenting the best day it ever had.
The second case covers twenty-six days, not the whole engagement. Daily tracking on that project began in July, so 1 August is the earliest clean baseline available. The honest framing is twenty-six measured days, and we will not stretch it into a longer claim just because the engagement started earlier.
The baseline and the current figure use the same keyword records. That matters more than it sounds. A portfolio can appear to double simply because keywords were added to tracking mid-window. We are not claiming portfolio-level growth on either account for exactly that reason, even though both grew, because the keyword sets changed and the comparison would not be clean.
Search volumes are monthly figures from the tracking platform for the specific geographic market named, not national totals.
What These Case Studies Do Not Claim
They do not claim revenue figures, because we do not have access to either client’s books and would not publish them if we did. They do not claim AI citation counts, because both engagements began before AI visibility measurement was part of the standard reporting layer; that measurement now runs monthly on new engagements and will appear in future case studies. They do not claim these results are typical, because results depend on the market, the starting condition of the site and how much verification cleanup the business needs first.
What they do claim is narrow and checkable: on these dates, in these markets, for these keywords, these positions were held by these pages.
See What the Rank Intelligence Engine Finds on Your Site
The first thing we would do for you is the same thing we did for both businesses above: measure where you actually stand against the companies beating you, then tell you what it would take. That read is free and takes thirty minutes.
Book your free strategy call, or start with an AI SEO audit if you would rather have the diagnosis in writing first.
Both businesses above are local service companies, and that work runs through our local AI SEO services. If you want the background first, read what AI search engine optimization is and how it works, or more about Phil Belleville, who led both engagements.