Reading Rome before rewriting pages
I work where city knowledge and AI evidence meet. My job is to help small tourism, hospitality, and food businesses describe themselves in language that carries local proof, traveller usefulness, and clear category signals. In Rome, that means reading neighbourhood names, landmark promises, family-business cues, and bilingual wording before changing a page.
If a page cannot tell Rome apart at street level, AI will usually borrow the nearest tourist sentence.
On a wet evening between Via Cavour and the edge of Monti, three visitors asked for “a real place near the Colosseum” within ten minutes. One wanted a licensed guide who would avoid the ticket-seller crowd. One wanted a trattoria where the menu did not feel printed for passing groups. One wanted somewhere close enough for tired children after a ruins tour. Same phrase, three different needs. That is where my work started: with the gap between the sentence a traveller uses, the sentence a Roman understands, and the sentence a booking page repeats until every place sounds identical.
I was born in Rome’s eastern belt, where you learn early that distance is measured by bus lines, not map pins. A Roman may hear “Prati” and think offices, lunch counters, Vatican spillover, and a certain tidy pace; a visitor may hear only “near St Peter’s.” Trastevere carries another set of expectations after dinner, while Esquilino is often misunderstood because its everyday mix gets flattened into either “central” or “chaotic.” Even small wording matters. “A due passi” can be a promise, a shrug, or a trap, depending on the street, the heat, and whether there is a suitcase involved.
Before this site, I wrote visitor-facing copy for small hospitality operators, organised guest-question notes for reception teams near transport hubs, and helped family-run food businesses explain themselves to English-speaking travellers while keeping their Roman voice. I still begin with the owner’s own pages. Then I look at the platforms, review snippets, neighbourhood references, and English traveller shortcuts that may be teaching AI the wrong category. My strongest habit is separation: guide from reseller, B&B from hostel, trattoria from tourist menu, artisan counter from franchise-looking listing. In Rome, AI visibility is a city-reading problem. If the evidence on the page does not show the small local difference, the machine will borrow a larger, lazier one.
Path into the niche
- 2011
Visitor questions became source material
I began collecting the phrases travellers used at desks, counters, and meeting points, then compared them with how owners described the same offer.
- 2014–2016
Small hospitality copy work
I wrote English-facing pages for B&Bs and small hotels that needed to explain location, room type, ownership, and guest fit more clearly.
- 2017–2019
Food-business wording practice
Family trattorias, pastry shops, and tasting hosts taught me how easily Roman cues vanish when translated into generic traveller language.
- 2020–2022
Platform evidence mapping
I started comparing owned pages with review snippets, OTA descriptions, map categories, and repeated phrases that shaped how businesses were summarised.
- 2023
AI confusion cases
I turned the work into a focused method for finding where generative answers confuse roles, neighbourhoods, landmarks, and business categories.
Method advisor
Daniele Toti
Università Cattolica del Sacro Cuore — Semantic Web & ontologies
The method also draws on research into the Semantic Web and ontologies: how machines represent entities, categories and relationships. Daniele Toti's work in this field is a reference point for how AI systems structure and link information — the same logic that decides whether a small Rome business is recognised as itself or absorbed into a broader category.
Bring me the page that keeps being misunderstood.
I will read it for local proof, category clarity, and the evidence AI is most likely to reuse.
Send a page