The Italian Cabinet registered theitaliancabinet.com on March 10, 2026 and launched the site the next day. A boutique studio in South Florida, it works directly with Italian manufacturers on handcrafted cabinetry, wardrobes and interior doors for high-end homes around Miami. The site had everything a customer would look for, product pages, the manufacturing partnerships, a way to book a showroom visit. What it had none of was history. No directory, review site or other page had ever named it, because it had existed on the web for one day.
That is the AI visibility most new businesses start with, and few ever measure it. When someone asked ChatGPT, Perplexity or Google AI about Italian cabinetry in Miami on March 11, the answers named Linea Studio, Pedini by Daruso and Scavolini, firms with years of web history behind them. The Italian Cabinet was in none of the fifteen answers. This case follows what happened next on the Reachd dashboard, week by week, and then what the same dashboard read four months after the owner stopped working on it.
Day one, and the business is not showing up in ChatGPT
The first Reachd analysis sent five customer questions to three assistants. Fifteen answers came back, and the business was in none of them. The firms the dashboard ranked above it were the same ones an owner in Miami would name from memory.
Those recommendations went to businesses with longer web histories on questions like “custom Italian cabinetry in Miami” and “where to find a showroom for Italian kitchen cabinets in South Florida.” A free AI visibility check shows the same first answer for any business, which is where this one started.
Why the assistants had nothing to say
The analysis found four gaps, and together they made The Italian Cabinet invisible to AI models:
- A brand new domain with no indexing history and no third-party mentions anywhere on the web
- No listings on business directories or review platforms that AI models use as trusted data sources
- No structured data on the website to help AI understand what the business offers and where it operates
- No independent citations confirming the business’s specialization or manufacturer partnerships
AI assistants like ChatGPT and Perplexity assemble their recommendations from patterns across many independent sources. A good website on its own rarely gives a model enough to recommend a business with confidence. The models look for external confirmation from directories, reviews and third-party mentions, and the reasons a business goes unrecommended are usually this short list.
The action plan for getting recommended by ChatGPT
Reachd produced a list of steps in priority order. At the top sat the same business details on every directory AI models treat as a trusted source, and manufacturer partnerships and specializations stated in a way a model can check. After that came structured data on the website and the outside mentions that give a model confidence in a recommendation.
Ann, the owner, started working through the highest-priority items that same day.
Named by ChatGPT and Perplexity within 48 hours
The analysis two days later found the business in two of the fifteen answers. Perplexity placed it second on the question about lead times, ChatGPT fourth on the one about fully custom cabinetry. Thirteen answers still had no mention.
Two answers is a small number, but both linked back to the website and described the business in terms an owner would recognize, which is the signal that an assistant has read something about a company rather than guessed at it. A domain two days old was being named.
On the short list for every question by March 16
The March 16 analysis ran a fresh set of five customer questions across the same three assistants, and every one of them returned The Italian Cabinet on at least one assistant.
On the question about what a fair price for real Italian cabinetry in Miami looks like, Perplexity returned The Italian Cabinet as the first result, ahead of every other firm it named. ChatGPT picked the business up on three separate questions, including the one about custom dimensions for an awkward condo layout and the one about telling authentic Italian manufacturing from marketing copy. Google AI named the business on none of the five, which is not unusual for a domain this young. Each assistant has its own sources and its own update cycle, and progress arrives unevenly across them.
A map of one day shows one day, since assistants regenerate their answers every time, and over the following two weeks the business held one or two of the five questions per analysis rather than all five. What March 16 proved was narrower and more useful, that the assistants had read the business and were willing to name it. Keeping a place on the short list took the next six weeks.
Six weeks in, recommended more often than the established showrooms
Through late March the dashboard read the business as present but behind. By mid-April it was named on most analyses, and on April 22 it moved ahead of the competitors the dashboard ranks it against, D’Cassa, European Cabinets and Spectrum Fine Cabinetry. On April 24 it was the most recommended of the four, six weeks after a start from nothing.
The dashboard reads the score over a week of analyses, so a single strong day never moves it far, and assistants regenerate their answers every time. Read over weeks rather than days, the direction is one way. The AI visibility score is the share of answers that recommend the business, and by late April that share was larger than any rival’s.
None of this involved a content campaign or backlink outreach. The changes were directory listings with matching details, structured data on the site, pages that state the manufacturer partnerships plainly, and the third-party mentions those bring.
Four months later, without touching the site
The April 24 analysis was the last one for a while. The owner stopped working on the site, and the analyses stopped with it. On August 24 Reachd measured the business again, on the same set of competitors, for the first time in four months, and it was still ahead of all of them.
A recommendation earned through the sources AI assistants trust does not expire when the work stops. It lasts as long as those sources do.
A directory profile, a page that states what the business is and a handful of independent mentions keep answering the assistants’ questions on their own, month after month. Paid placement stops the day the budget does. The dashboard in August read a business the assistants had settled on, and nobody had done anything about it since April.
What to take away from this
A domain with no history is not the disadvantage it looks like, provided the work is aimed at what assistants read. The Italian Cabinet went from absent in every answer to named by ChatGPT and Perplexity in two days, to the most recommended of its Miami competitors in six weeks, and stayed there for four months with no further effort.
Which assistants responded, and how, matters too. ChatGPT and Perplexity surface businesses differently from traditional search. Instead of ten pages of links, they name a handful of options and explain why each one deserves attention, so a business that makes that short list gets a disproportionate share of the conversation. The same pattern holds in other Miami categories, as the ChatGPT realtor recommendations in Miami dataset shows.
For businesses in categories like luxury home services or interior design, where customers do careful research before reaching out, it may be worth checking whether AI assistants are even aware they exist, which is where AI search optimization starts. For this one, on March 11, they were not.
ChatGPT 
Perplexity