GEO for businesses: AI visibility by industry and by type of organization

A Refractia scan follows the same method in every sector: it reads your website, builds the brand profile (category, market, traits, and competitors you can edit), and asks the five engines category, comparison and purchase-intent questions, in the market's language and with the country named in full.

• Every rate comes with its sample (k of n). A 25% with no denominator can't tell 1 of 4 answers apart from 250 of 1,000.
• Not measured is not zero. If an engine didn't answer, that question doesn't count as "answered and didn't name you."
• Chat and AI search, reported separately. Chat answers from what the model knows; AI search paraphrases what ranks on the web.

Scans are started by a person: Refractia is a deep diagnostic, not a daily tracker.

For each industry, a typical question people ask AI before they decide, and the measurement that tells you the most. The questions are generic examples: in a real scan they are generated for your category and your market, in the market's language and with the country named in full.

Banking & financial services

People ask "What bank should I use for my small business in the United States?" and get a short list with a one-line reason next to each name. That line is where the risk sits: an old fee, a discontinued product or a past problem can still be how an engine describes you. AI Perception shows the traits each engine attaches to your bank (fees, service, trust) and the sources those descriptions come from.

Insurance

"What's the best car insurance for young drivers in Colombia?" is a question about segments: the right answer depends on who is asking. Engines also describe how each carrier handles claims, often from forums and review sites rather than from your policy pages. Performance by Topic puts you against each rival on price, claims and service, so you can see for which customers you are recommended.

Healthcare

Patients ask "What's the best hospital for cardiac care in Buenos Aires, Argentina?", and physicians ask similar questions before a referral. Accuracy matters more than mentions here: an engine that lists a specialty you closed, or misses one you opened, sends patients elsewhere. Refractia checks what each engine says about your services and which pages it cites (Influence Sources), so a wrong description can be traced back to where it came from.

Retail & e-commerce

"Where should I buy sustainable clothing online in Mexico?" returns a handful of stores, and for many shoppers the search starts and ends with that list. We measure your Share of Voice on shopping questions and, if you sell online, how often AI names your store and where it sends the click (AI Commerce).

Consumer packaged goods

For packaged goods the typical question is a purchase question: "Which laundry detergent goes furthest in Chile?" A brand that leads on the shelf can still be missing from the answer. The scan counts how often you make the short list, in what position, which brands take your place when you don't, and whether engines recommend your other lines or only the flagship.

Food & beverage

Questions mix product and occasion: "Which non-alcoholic beer can I find in supermarkets in Spain?", or which brand suits a gluten-free diet. Engines sort brands by taste, health and price, and they can only recommend what they know you make. Incidence by topic shows on which of those axes you are named and on which a rival is.

Energy & utilities

"What's the best gasoline in Argentina?" comes with follow-ups on price, rewards programs and sustainability, and AI answers each one differently. Sustainability claims in particular tend to be described from press coverage rather than from your own reports. Position and sentiment are measured topic by topic, which shows whether a strong reputation for quality hides a weak one on price.

Automotive

"What's the most reliable compact SUV in Mexico?" is answered as a comparison on safety, fuel economy and after-sales service, and buyers ask it for their situation: a family, a commute, a first electric car. Chat engines answer from what the model already knows; AI search paraphrases current reviews. We report the two separately, because a recent model often appears in AI search before chat engines know it exists.

Technology & SaaS

Software buyers ask for comparisons, such as "Which CRM is easiest to use for a small team in the United States?", and engines sum up each product in one line. If your differentiator isn't clear on your own pages, you get a generic line or none. Refractia measures your share of voice in comparison questions and how readable your pages are to AI crawlers (GEO Optimization).

Telecommunications

"Which carrier has the best mobile coverage in Colombia?" is answered from coverage maps, press and user complaints, and the description can lag years behind your network. Sentiment is tracked by topic (coverage, price, customer service) together with its evolution from one scan to the next, so you can tell whether an upgrade has reached what AI says.

Hospitality & travel

Travelers ask for a property that fits the trip: "What's the best hotel in Cancún, Mexico, for a family with kids?" Whether an engine calls you family-friendly or a romantic getaway decides who considers booking, and those labels come mostly from travel and review sites. The scan shows which of those sites each engine cites, market by market.

Education

Students and parents ask "Which university has a strong computer science program in Chile?", and the answer blends rankings, job outcomes and cost, so AI's view of your standing can differ from the official rankings. We check whether you are named for the programs you lead in, and with which adjectives: prestige, job prospects, cost.

Real estate

"Which developer is reliable for buying off-plan in Argentina?" is a question about trust before it is about price. Engines build the answer from listing portals, press and reviews. Refractia checks whether reliability shows up among the traits AI gives you, and which portals it leans on.

Media & entertainment

Readers ask engines which outlet to trust: "Which newspaper is most reliable for economic news in Mexico?" For a media brand, the way AI describes its credibility is part of the product. You see how each engine describes your outlet and which sources it relies on when it does.

Logistics

Store owners and procurement teams ask "Which shipping company is best for an online store in Argentina?" and treat the answer as a first vendor list. Engines also need to know your full range (freight, warehousing, last-mile, cross-border); a provider known for one service is left out of the rest. The scan measures share of voice in these B2B questions and your position on cost, delivery times and coverage.

Professional services

Clients ask for a specialty and a city: "Which law firm is good at employment law in Bogotá, Colombia?" A well-known firm can go unnamed because the engine doesn't connect it with that practice area. We check whether AI knows what you specialize in, and whether it names you when the question doesn't mention you.

Luxury & fashion

Luxury questions are about rank and lasting value: "Which watch brand holds its resale value best in Mexico?" Engines rank brands and retell their history, and a description that drops craftsmanship or exclusivity flattens a house into its category. Refractia measures alignment with the traits you declare and which specialist media the engines cite.

Nonprofits

Donors ask where to give: "Which environmental nonprofit has a good reputation in Argentina?" The answer is a few names and a judgment on impact and transparency. You see whether you are named when someone wants to support your cause, and how engines describe your transparency and results.

• In trust industries (banking, insurance, healthcare), an error in how AI describes you weighs more than a point of share of voice: look at descriptors, sentiment and sources first.
• In consumer goods, retail and food, the typical question is a purchase question and AI returns a short list: what matters is how often you make the list and in what position.
• In B2B and professional services, a brand can be missing because AI doesn't know what it does: start with what your website says about you.
• In brands with several product lines (for example, an electronics brand that sells phones and air conditioners), each line has different rivals and is measured separately (Product lines).

Brands in energy, banking, consumer goods and retail already work with Refractia, among them Pan American Energy, AXION energy, Bancor, Grido, Sendbox, Universal, Bando, Titan Push and Castrol.

Plans are quoted: FUNDACIONAL for small businesses and challenger brands, CRECIMIENTO for scaleups, and ENTERPRISE for large brands. There are no published prices; the initial diagnostic is free.