Found Everywhere the Question Gets Asked: How Uma Marisqueira Macau Became the #1 Seafood Recommendation on Google, AI Mode and ChatGPT

One query - “Where to eat the best seafood restaurant in Macau?” - asked of three different systems on the same day returned the same business first in all three. Google Search placed Uma Marisqueria Macau at the top of the local Places pack. Google’s AI Mode named it the lead pick in its generated answer, with a full rich-result card. ChatGPT recommended it above every other Macau seafood restaurant it listed. Underneath that, the Google Business Profile logged 46,709 views from 24,426 searches between March and August 2026, with 72% of all discovery coming from Google Maps on mobile.

Uma Marisqueira Macau - 正宗葡式海鮮飯The Sovereign of Authentic Portuguese Seafood in Taipa - Arroz de Marisco, Portuguese Seafood Rice, Authentic Portuguese, Taipa Village, Macau Seafood, Fresh Seafood, Polvo à Lagareiro, Grilled Sardines, Amêijoas à Bulhão Pato, Lisbon Legacy, Old Taipa, Macanese Flair, Family Restaurant, Traditional Seafood, Crab Curry, Seafood Pot, Macau Dining, Uma Marisqueira, Heritage Cuisine, Authentic Flavors, Uma Marisqueria Macau, Uma Marisqueira Macau, Authentic Portuguese food Taipa, Best seafood rice Umarella

Why this case matters more than a ranking screenshot

Most local SEO case studies show a position and stop. This one is worth reading for a different reason: it is a controlled look at what happens when the same question is asked of three systems that work in completely different ways.

Google Search ranks business entities by relevance, distance and prominence. Google’s AI Mode generates prose and chooses which businesses to name inside it. ChatGPT draws on an entirely separate corpus and returns its own shortlist. These are three independent judgements about the same question, and they agreed.

That agreement is the finding. It suggests the underlying signals category accuracy, review quality, consistent machine-readable business data are being read similarly by systems that share almost no infrastructure.

Table of Contents

1. Business snapshot

2. The challenge: search fragmented into three

3. The strategy: SEO, GEO and AIO together

4. The results across three engines

5. The profile data behind the rankings

6. What the data actually shows

7. The competitive context

8. How this is replicable

9. What we would do next

10. Methodology and honest caveats

11. Frequently asked questions

12. The 500+ semantic keyword bank

1. Business Snapshot

The listing name carries a Chinese descriptor 正宗葡式海鮮飯, roughly “authentic Portuguese-style seafood rice” alongside the Latin-script name. In a market where a substantial share of searches are conducted in Chinese, that descriptor does work that an English-only name cannot.

2. The Challenge: Search Fragmented Into Three

2.1 The old problem and the new one

For twenty years, being findable meant one thing: rank on Google. A restaurant either appeared in the local pack or it did not, and the work was to appear there.

That is now the first of three problems.

2.2 How a diner actually searches in 2026

The same person looking for seafood in Macau might:

• Type into Google and scan the map pack

• Ask Google’s AI Mode for a recommendation and read the generated answer

• Ask ChatGPT directly, conversationally, and take the shortlist it gives back

2.2.1 Why winning one does not win the others

These systems fail independently. A profile can rank first in the local pack and be entirely absent from the AI-generated answer sitting above it because the AI is drawing on different signals, in a different order, with different tolerance for ambiguity.

The specific failure modes:

The asymmetry that catches businesses out

You get no notification when an AI system declines to mention you. Rankings hold, the profile looks healthy, and the business simply stops appearing in the answers a growing share of customers rely on. There is no Search Console message for “not recommended.”

3. The Strategy: SEO, GEO and AIO Together

The work ran across three layers simultaneously rather than sequentially. That sequencing decision is the substance of this case study.

3.1 SEO - winning the classic local pack

Conventional local search optimisation, done completely rather than partially.

3.1.1 What was done

• Primary category set precisely to Seafood restaurant rather than the generic Restaurant, with Portuguese as a supporting category

• Business name aligned with the real trading name including its Chinese descriptor, so the listing matches how customers actually refer to and search for it

• Attributes completed in full - service options, dining options, accessibility, payment methods

• Price range published (MOP 150–300), which is both a ranking input and a filter customers use

• Review velocity established - a steady flow of recent reviews rather than a historic pile

• Photography refreshed across food, interior and exterior

Why the category matters most

Primary category is closer to a switch than a dial. It determines which queries the listing is eligible for at all. A seafood restaurant filed under “Restaurant” competes against every eating establishment in Macau; filed under “Seafood restaurant” it competes against seafood restaurants and still appears for general restaurant searches. That single field is the highest-leverage change available in local search, and it is free.

3.2 GEO - being chosen inside generated answers

Generative Engine Optimisation: structuring the same underlying information so a system writing prose can lift it out with confidence.

3.2.1 What that required

• Unambiguous category language. An AI writing “Authentic Portuguese Marisqueiras” as a heading needs to be certain the business belongs under it. Ambiguity is resolved by omission.

• Complete, current factual fields. Rating, review count, price band, hours, neighbourhood these are the components of the rich card AI Mode rendered. Missing fields mean a thinner card or no card.

• Review content that describes the offer. Reviews naming the seafood rice and the grilled sardines give a generative system concrete, attributable detail to summarise. Reviews saying only “great place” give it nothing.

• Consistency between what the profile claims and what reviews corroborate. Where those two disagree, systems default to caution.

3.3 AIO - being recommended by third-party assistants

ChatGPT does not read your Google Business Profile the way Google does. It draws on a separate corpus and its own retrieval. The work here is making the business a confidently resolvable entity across the wider web.

3.3.1 What that meant in practice

• One canonical name, used identically everywhere

• Consistent address and location descriptors across every surface that mentions the business

• Category, price band and rating consistent wherever they appear, so nothing contradicts

• Third-party corroboration the business described by sources other than itself

The principle underneath all three layers

None of this is three separate jobs. It is one job making accurate, complete, consistent information about a business available in machine-readable form evaluated by three different systems. The reason the layers are named separately is that each one fails separately, and you cannot tell which is broken without checking all three.

4. The Results Across Three Engines

Query tested: “Where to eat the best seafood restaurant in Macau?” Location: Macau Date: As captured, August 2026

4.1 Google Search first in the Places pack

The local pack returned three businesses. Uma Marisqueria Macau placed first.

Below the pack, umarella.org’s own seafood restaurant guide appeared in the organic results for the same query.

4.2 Google AI Mode the lead recommendation

AI Mode generated a structured answer opening with the observation that Macau offers seafood ranging from authentic Portuguese marisqueiras to traditional Cantonese establishments. Under the heading “Authentic Portuguese Marisqueiras”, it named Uma Marisqueria Macau as the top choice.

The listing rendered as a full rich card inside the generated answer:

• Rating and review count (4.8, 265)

• Price band (MOP 150–300) and category

• Open status and neighbourhood (Freguesia de Nossa Senhora do Carmo)

• Call, Directions and Website buttons - actionable, inside the AI answer itself

The answer described it as the top choice for a genuine taste of Portugal in Macau, noting its relationship to the Lisbon operation and praising its signature seafood rice.

4.2.1 Why the rich card is the significant detail

Being mentioned in an AI answer is worth something. Being rendered as a card with a call button is worth considerably more, because the conversion path never leaves the answer. The card is only possible when the underlying structured data is complete rating, price, hours, location, contact. Incomplete profiles get a text mention at best.

4.3 ChatGPT - top of the shortlist

Asked the same question, ChatGPT returned a map with pinned recommendations and cards beneath it. Uma Marisqueria Macau appeared first, ahead of Fernando’s Restaurant (4.3 ★, MOP 200–450), with rating, category, price band and open status displayed.

The methodological point worth noting

This response was generated in a logged-out session. There was no account history, no saved preferences and no personalisation to influence the result. Whatever produced that ranking was the model’s own retrieval and judgement, not a reflection of the user.

4.4 Three systems, one answer

5. The Profile Data Behind the Rankings

Rankings are a snapshot. Profile performance over six months is the substance. All figures below are from Google Business Profile performance reporting for March–August 2026, with August partial at the time of capture.

5.1 Discovery volume

5.2 Platform and device breakdown

Derived from those figures:

• Mobile accounts for 43,289 views - 92.7% of all discovery

• Maps accounts for 34,536 views - 73.9%, against 26.1% for Search

• Desktop across both surfaces totals 3,420 views, 7.3%

5.3 The search terms that surfaced the profile

These five account for approximately 21,078 of 24,426 searches - 86.3% of all searches that surfaced the profile.

A note on the “17k” figure

Google’s interface rounds larger numbers. The true value sits somewhere between 16,500 and 17,499. We have written it as approximately 17,000 rather than treating a rounded display figure as precise a distinction worth maintaining in any reporting you rely on commercially.

6. What the Data Actually Shows

Anyone can present a dashboard. The value is in what the numbers mean operationally.

6.1 Non-branded discovery outnumbers branded by roughly 17 to 1

This is the most important figure on the page.

Within the top five search terms, non-branded, category-level searches account for approximately 19,901 searches “restaurants”, “restaurantes”, “portuguese restaurant” against 1,177 branded searches for the business by name.

That is a ratio of about 16.9 to 1.

6.1.1 Why this is the number that matters

Branded search traffic means people who already knew the business looked it up. That is reputation, word of mouth and prior visits arriving through the search box. It is valuable, and it is not evidence that local SEO is working.

Non-branded discovery means people who did not know the business existed found it while searching for a category. That is the entire point of local search, and it is the only number that demonstrates the work produced new customers rather than servicing existing ones.

The diagnostic to apply to your own profile

Open your Business Profile performance report and split searches into branded and non-branded. If the majority are branded, your profile is a lookup destination, not a discovery engine and the opportunity in front of you is large.

6.2 Google Maps on mobile is not a channel, it is the channel

At 72% of all discovery - 33,612 views - Maps on mobile dominates everything else combined.

6.2.1 What follows operationally

• Photography is the storefront. On a mobile Maps card, images occupy more visual space than text. Poor photos lose the customer before any copy is read.

• The “Directions” action is the conversion. Someone requesting directions on a phone is physically going. That is closer to a booking than a website click.

• Attributes function as filters. On mobile Maps, users filter by open now, price, rating and features. A missing attribute removes you from filtered results entirely not lowers you, removes you.

• Desktop optimisation is a rounding error here. At 7.3% combined, desktop deserves proportionate attention and no more.

6.3 The profile serves at least two language markets simultaneously

The search terms reveal three linguistic entry points:

• “restaurants” - English, ~17,000 searches

• “restaurantes” - Portuguese, 2,434 searches

• Chinese-script branded queries including 正宗葡式海鮮飯

In Macau a market with Portuguese heritage, a Cantonese-speaking majority and heavy international tourism this multilingual surfacing is not incidental. It is the direct result of a listing name and profile that carry meaning in more than one language.

6.3.1 The transferable lesson

Where a business genuinely operates across languages, the profile should reflect that. Not through keyword stuffing, but through the real name, accurate categories and reviews written by customers in their own languages.

6.4 A caution about the two headline metrics

46,709 views against 24,426 searches produces a ratio of 1.91. It is tempting to read that as engagement roughly two views per qualifying search.

Do not. Google defines these two metrics differently, and they count different events across different surfaces. They are not a numerator and denominator, and treating the ratio as a click-through rate would be a reporting error. We include both because both are reported; we decline to invent a relationship between them.

Why we are flagging this rather than using it

Because a case study that manufactures an impressive-sounding derived metric is a case study you should distrust on everything else. The real numbers are strong enough.

7. The Competitive Context

Ranking first is meaningful only relative to what it beat.

7.1 Beating a listing with roughly eight times the reviews

Portucau 澳葡坊 holds the same 4.8 rating from approximately 2,000 reviews. Uma Marisqueria Macau ranks above it with 265.

This is the clearest available demonstration that review count alone does not determine local ranking. If it did, the result would be reversed and no amount of optimisation would change it.

7.1.1 What plausibly explains the gap

• Category precision. Uma Marisqueria is categorised as Seafood; Portucau as Portuguese. For a query explicitly asking about seafood, the more precise category match is the stronger relevance signal.

• Review recency and response. Post-2026 core update analysis indicates Google weights recent review flow and owner response rate more heavily than accumulated volume.

• Profile completeness. A fully populated profile attributes, price, photos, hours is an explicit input.

• Proximity, which varies by searcher and is discussed in the caveats below.

7.2 The third result proves the point in reverse

Gold Coast Oyster and Seafood Restaurant holds a 5.0 rating from a single review and placed third. A perfect average from one review carries almost no weight. Google is clearly evaluating something considerably more sophisticated than star ratings.

7.3 The ChatGPT comparison

ChatGPT ranked Uma Marisqueria (4.8) above Fernando’s Restaurant (4.3) a genuinely well-known Macau establishment with substantial independent coverage. That an assistant drawing on a separate corpus reached the same conclusion as Google suggests the underlying entity signals are consistent enough to be read the same way twice.

The honest reading

This is one query on one day. It is evidence, not proof. Section 10 sets out precisely what it does and does not establish.

8. How This Is Replicable

Nothing in this case study required a media budget, a website rebuild or a proprietary tool. Here is the sequence, in the order it produces results.

8.1 The foundation - weeks one to two

8.1.1 Fix the primary category

Search your main category term from your location. Look at the three businesses in the local pack. Check what primary category each uses. Choose the most precise description of your core offering, not the broadest.

8.1.2 Complete every field

Hours including holidays, full address, price range, phone, website, description, services, products.

8.1.3 Populate every applicable attribute

These function as filters on mobile Maps. Missing attributes remove you from filtered searches.

8.1.4 Add real photography

Food, interior, exterior, and the entrance so people can find the door. Mobile Maps is a visual surface.

8.2 The signals - weeks three to twelve

8.2.1 Establish review velocity

Ask every satisfied customer, in person, at the moment of satisfaction. Never gate requests to happy customers only, never incentivise. Both are platform violations and both are unlawful under EU consumer rules.

8.2.2 Respond to every review

Within 48 hours, in the language it was written in. Response rate is now a weighted signal, and the audience for a reply to criticism is everyone reading it afterwards.

8.2.3 Post consistently

Weekly is a reasonable cadence. It occupies space in your knowledge panel that would otherwise be empty.

8.3 The consistency layer - ongoing

8.3.1 One canonical name everywhere

Identical spelling, identical formatting, on every platform that mentions you.

8.3.2 Identical address and phone

Character for character. Entity resolution sharpened in 2026; inconsistency costs more than it used to.

8.3.3 Claim every relevant third-party profile

Apple Business Connect, Bing Places, sector platforms, review sites. These are where assistants corroborate you.

8.4 What to measure

The monthly AI check, specifically

Once a month, ask Google AI Mode, ChatGPT and one other assistant the question a customer would ask to find a business like yours. Record whether you appear and in what position. It takes four minutes and it is currently the only reliable way to measure this.

9. What We Would Do Next

A case study that presents a finished job is usually presenting a stopped job. What follows for this listing:

• Broaden non-branded coverage beyond “restaurants” into dish-level and occasion-level terms — seafood rice, grilled sardines, group dining, waterfront dining

• Expand Chinese-language review volume, since Chinese-script queries are already surfacing the profile

• Build third-party corroboration on travel and review platforms, which is what strengthens assistant recommendations specifically

• Track AI visibility systematically rather than through periodic manual checks

• Convert Maps visibility into bookings, since direction requests and calls matter more than views

• Monitor competitors’ review velocity, because a 265-review lead over a 2,000-review competitor rests on recency and precision that must be maintained

10. Methodology and Honest Caveats

Every case study should include this section. Most do not, which is itself informative.

10.1 What the ranking results are

Three searches for one query, from one location, on one day, captured as screenshots. They demonstrate that the business ranked first in all three systems at that moment.

10.2 What they are not

10.2.1 Local rankings vary by the searcher’s exact position

Distance is a ranking factor in local search. A result captured in Macau does not mean a searcher three kilometres away sees the same order. Honest local reporting uses geo-grid sampling across many points; a single capture shows one vantage point. This is a real limitation and we state it rather than let a screenshot imply universal coverage.

10.2.2 AI answers are non-deterministic

Generated responses vary between sessions, phrasings and users. The same question asked tomorrow may produce a different ordering. What is durable is the underlying signal quality; the specific output is a sample of it.

10.2.3 Correlation, not proof of causation

The work was done and the results followed. That is strong circumstantial evidence, and it is not a controlled experiment. There is no counterfactual version of this restaurant that did nothing, and anyone claiming otherwise about any case study is overstating.

10.2.4 The August data is partial

The reporting period runs March to August 2026, captured mid-August. The final month is incomplete, so totals understate the full six-month period slightly.

10.3 What is verifiable

• The profile figures are Google’s own reporting, unmodified

• The platform breakdown sums exactly to the stated total of 46,709

• The ranking screenshots are unedited captures

• The ChatGPT result was generated in a logged-out session with no personalisation

Why we publish the caveats

Because a prospective client evaluating an agency should be able to tell the difference between evidence and marketing. A case study that survives its own caveats is worth more than one that avoids them — and an agency willing to tell you the limits of its own results is more likely to tell you the truth about yours.

11. Frequently Asked Questions

11.1 What is the difference between SEO, GEO and AIO?

SEO wins classic ranked results. GEO (Generative Engine Optimisation) is being selected inside AI-generated answers such as Google’s AI Mode. AIO covers being recommended by third-party assistants like ChatGPT. Same underlying information, three systems reading it.

11.2 Do I need to do all three separately?

Largely no — they share a foundation of accurate, complete, consistent, machine-readable business data. But they fail separately, so all three need checking independently.

11.3 How long did this take?

The profile data covers March to August 2026. Google Business Profile changes typically show movement within two to eight weeks; broader local consolidation takes three to four months.

11.4 Does this work outside Macau?

The mechanics are the same anywhere Google operates. Competitive density and language mix change the difficulty, not the method.

11.5 Can a business with fewer reviews outrank one with thousands?

This case demonstrates it can — 265 reviews ranking above approximately 2,000. Category precision, review recency, response rate and profile completeness all contribute.

11.6 How do I know if AI systems recommend my business?

Ask them. Once a month, put your category question to Google AI Mode, ChatGPT and one other assistant and record whether you appear. Tooling for this is immature; manual checking is currently more reliable.

11.7 Why does Google Maps mobile dominate so heavily?

Because local searches happen on phones, usually near the moment of deciding. For this listing it was 72% of all discovery. Most businesses under-invest in the surface where nearly three quarters of their visibility occurs.

11.8 What is the single highest-impact change?

The primary category. It determines which searches you are eligible for at all, it takes thirty seconds to change, and a surprising number of businesses have it wrong.

11.9 Is adding descriptive words to a business name allowed?

Only if it is genuinely the business’s real, publicly used name — what appears on the signage and documentation. Adding keywords you do not actually trade under is the leading cause of listing suspension.

11.10 Can you guarantee these results?

No, and neither can anyone else. Rankings depend on competition, proximity, category density and factors outside any agency’s control. What is repeatable is the method; the outcome varies by market.

11.11 What does a programme like this cost?

Local profile management typically runs 300€–800€ per month for a single location. Our companion guide on hiring an SEO agency sets out what each deliverable should cost and how to verify it was done.

11.12 What should I do first?

Open your Business Profile. Check the primary category, the attributes, the photo count and the oldest unanswered review. That hour will tell you more about your position than any audit you could buy this week.

12. The 500+ Semantic Keyword Bank

Built for case study, proof-of-results and AI visibility intent — the terms used by people evaluating whether this kind of work produces outcomes, and by businesses searching for the problems it solves. PT-PT vocabulary in the Portuguese clusters.

12.1 SEO, GEO and AIO terminology

• SEO GEO AIO

• generative engine optimisation

• GEO optimisation

• answer engine optimisation

• AEO vs GEO

• AI optimisation for businesses

• AI visibility optimisation

• LLM optimisation

• AI search optimisation

• optimising for AI Mode

• optimising for ChatGPT

• optimising for Perplexity

• optimising for Gemini

• AI recommendation optimisation

• being recommended by AI

• getting cited by AI

• AI answer visibility

• generative search visibility

• conversational search optimisation

• assistant recommendation ranking

• multi engine visibility

• cross platform search visibility

• three layer search strategy

• otimização para motores generativos

• otimização para respostas de IA

• visibilidade em IA

• ser recomendado pela IA

• ser citado por assistentes

• otimização para ChatGPT

• pesquisa conversacional

12.2 Case study and proof-of-results terms

• SEO case study

• local SEO case study

• Google Business Profile case study

• restaurant SEO case study

• AI visibility case study

• GEO case study

• before and after SEO results

• SEO results example

• proven SEO results

• documented SEO outcomes

• local SEO success story

• restaurant marketing case study

• hospitality SEO results

• measurable SEO outcomes

• SEO ROI example

• agency portfolio results

• verified agency results

• client results local search

• ranking improvement case study

• profile views growth

• estudo de caso SEO

• estudo de caso SEO local

• resultados comprovados SEO

• caso de sucesso marketing local

• resultados de perfil de empresa

12.3 Google Business Profile performance

• Google Business Profile views

• business profile performance report

• profile discovery metrics

• searches showed your profile

• platform and device breakdown

• Google Maps mobile views

• Google Search mobile views

• desktop versus mobile discovery

• direction requests metric

• calls from Business Profile

• website clicks from profile

• bookings from Google

• menu views metric

• search terms report GBP

• branded versus non branded searches

• discovery searches

• direct searches

• customer actions per month

• profile insights explained

• how to read GBP performance

• visualizações do perfil de empresa

• relatório de desempenho do perfil

• pedidos de direções

• chamadas a partir do perfil

• termos de pesquisa do perfil

• pesquisas de marca versus sem marca

12.4 Local pack and map ranking

• local pack ranking

• map pack first position

• ranking number one Google Maps

• Google Maps ranking factors

• local pack three results

• Places pack Google

• proximity ranking factor

• relevance distance prominence

• geo grid rank tracking

• local visibility radius

• outranking competitors locally

• beating higher review competitors

• primary category ranking impact

• category selection local SEO

• secondary categories strategy

• attributes as map filters

• profile completeness ranking

• review recency weighting

• owner response rate signal

• classificação no pacote local

• primeira posição no Google Maps

• fatores de ranking local

• proximidade como fator

• categoria principal impacto

• atributos como filtros

12.5 Restaurant and hospitality search

• best seafood restaurant near me

• where to eat seafood

• seafood restaurant Macau

• Portuguese restaurant Macau

• marisqueira Macau

• Portuguese seafood restaurant

• authentic Portuguese dining

• arroz de marisco

• seafood rice restaurant

• grilled sardines restaurant

• Macanese Portuguese cuisine

• best restaurants Macau

• restaurant recommendations Macau

• Taipa restaurants

• Coloane dining

• Macau dining guide

• Macau food scene

• Portuguese food Asia

• restaurante de marisco

• melhor marisqueira

• arroz de marisco restaurante

• cozinha portuguesa em Macau

• restaurantes portugueses na Ásia

• guia de restaurantes Macau

12.6 Restaurant marketing and visibility

• restaurant local SEO

• restaurant Google Business Profile

• restaurant map pack ranking

• restaurant review generation

• restaurant photo optimisation

• menu on Google Business Profile

• restaurant reservation integration

• restaurant marketing agency

• hospitality digital marketing

• restaurant visibility online

• getting more diners from Google

• restaurant discovery searches

• restaurant category optimisation

• restaurant attributes Google

• restaurant posting strategy

• restaurant review responses

• marketing para restaurantes

• SEO local para restaurantes

• perfil de empresa para restaurantes

• avaliações para restaurantes

• fotos de restaurante Google

• ementa no perfil Google

• aumentar clientes através do Google

12.7 Multilingual and multi-market search

• multilingual business profile

• Chinese language business listing

• Portuguese language search terms

• bilingual restaurant listing

• multi language reviews

• Cantonese search optimisation

• Chinese script business name

• localised business name

• language mix local search

• tourist versus local searches

• international visitor discovery

• cross language category terms

• restaurants versus restaurantes

• translated menu SEO

• multilingual attributes

• perfil de empresa multilingue

• pesquisas em várias línguas

• termos de pesquisa em português

• avaliações em várias línguas

• nome de empresa em caracteres chineses

• turistas versus residentes pesquisa

12.8 AI Mode and generative answers

• Google AI Mode

• AI Mode local results

• AI Overviews restaurants

• generated answer recommendations

• rich result card AI answer

• AI answer business card

• call button in AI answer

• AI Mode citations

• AI Mode ranking

• zero click local search

• AI answer above local pack

• generative answer local intent

• AI shortlist recommendations

• how AI picks businesses

• why AI does not mention my business

• AI Mode versus classic search

• modo IA do Google

• respostas geradas por IA

• resultados locais em IA

• cartão de resultado em resposta de IA

• porque a IA não menciona o meu negócio

12.9 ChatGPT and assistant recommendations

• ChatGPT restaurant recommendations

• ChatGPT local search

• ChatGPT maps results

• asking ChatGPT where to eat

• assistant business recommendations

• ChatGPT shortlist businesses

• appearing in ChatGPT answers

• ChatGPT logged out results

• personalisation in AI answers

• Perplexity local recommendations

• Gemini business recommendations

• AI assistant discovery

• conversational restaurant search

• voice assistant recommendations

• recomendações do ChatGPT

• aparecer nas respostas do ChatGPT

• pesquisa local em assistentes

• descoberta através de IA

12.10 Entity consistency and machine readability

• entity consistency

• business entity resolution

• consistent business name

• NAP consistency

• machine readable business data

• structured business information

• schema markup restaurant

• LocalBusiness schema

• Restaurant schema type

• price range markup

• opening hours markup

• aggregate rating display

• knowledge graph entity

• knowledge panel accuracy

• third party corroboration

• data consistency across platforms

• canonical business name

• listing accuracy audit

• consistência de entidade

• dados legíveis por máquina

• consistência entre plataformas

• nome canónico da empresa

• auditoria de precisão de listagens

12.11 Reviews and reputation signals

• review velocity

• review recency

• review response rate

• responding to reviews

• review count versus quality

• outranking more reviewed competitors

• review content keywords

• reviews describing dishes

• multilingual review responses

• review generation ethics

• review gating prohibited

• incentivised reviews policy

• fake review risk

• star rating versus ranking

• one review five stars

• velocidade de avaliações

• recência das avaliações

• taxa de resposta a avaliações

• responder em várias línguas

• avaliações que descrevem pratos

• filtragem de avaliações proibida

12.12 Measurement and reporting

• local SEO reporting

• honest SEO reporting

• case study methodology

• verifiable results agency

• geo grid sampling

• ranking snapshot limitations

• non deterministic AI results

• correlation not causation SEO

• baseline and after comparison

• unmodified platform data

• screenshot evidence limitations

• monthly AI visibility check

• tracking AI citations

• KPI selection local business

• direction requests as conversion

• calls as conversion metric

• relatórios de SEO local

• metodologia de estudo de caso

• limitações de capturas de ecrã

• verificação mensal de visibilidade em IA

• indicadores para negócios locais

12.13 Question-based long-tail

• how do I rank first on Google Maps

• how do I get recommended by ChatGPT

• how do I appear in AI Mode answers

• why does my competitor rank higher with fewer reviews

• can fewer reviews outrank more reviews

• what is the most important Google Business Profile field

• how long does local SEO take to work

• how do I know if AI recommends my business

• what percentage of local searches are on mobile

• how do I measure Google Business Profile performance

• what are branded versus non branded searches

• how do I optimise for three search engines at once

• does ChatGPT use Google Business Profile data

• how do I get a rich card in AI answers

• can I add keywords to my business name

• como aparecer em primeiro no Google Maps

• como ser recomendado pelo ChatGPT

• como aparecer nas respostas de IA

• porque o concorrente aparece à frente

• quanto tempo demora o SEO local

• qual o campo mais importante do perfil

• como medir o desempenho do perfil

12.14 Commercial intent

• hire local SEO agency

• Google Business Profile management service

• AI visibility service

• GEO agency

• restaurant SEO agency

• request a free profile audit

• local SEO audit free

• book a strategy call SEO

• agency with proven results

• see agency case studies

• local SEO pricing

• GBP management cost

• contratar agência de SEO local

• gestão de perfil de empresa

• auditoria gratuita do perfil

• serviço de visibilidade em IA

• agência com resultados comprovados

• preço de SEO local

12.15 Semantic support and entity terms

• local search ecosystem

• discovery surface

• search fragmentation

• multi channel discovery

• micro moments local

• near me intent

• implicit local intent

• mobile first behaviour

• conversion path local

• footfall attribution

• store visit intent

• category level searches

• head terms versus long tail

• search demand by language

• competitive density local

• market saturation restaurants

• first mover advantage local search

• profile as storefront

• visual first discovery

• ecossistema de pesquisa local

• fragmentação da pesquisa

• intenção local implícita

• comportamento mobile first

• pesquisas ao nível da categoria

• densidade competitiva local

12.15b Problems this work solves

• my business does not appear on Google Maps

• not showing in local pack

• lost visibility in local search

• competitor outranking me locally

• profile views dropped

• no calls from Google Business Profile

• few direction requests

• profile only found by name

• no new customers from Google

• invisible in AI answers

• ChatGPT recommends competitors not me

• AI Mode does not mention my business

• not cited by AI assistants

• listing suspended Google

• wrong category on my listing

• duplicate listings same business

• inconsistent business information online

• old address still showing

• reviews not increasing

• reviews without responses

• poor photos on listing

• menu not showing on Google

• attributes missing from profile

• filtered out of Maps searches

• o meu negocio nao aparece no Google Maps

• nao apareco no pacote local

• perdi visibilidade local

• concorrente aparece a frente

• poucas chamadas do perfil

• so me encontram pelo nome

• invisivel nas respostas de IA

• o ChatGPT recomenda concorrentes

• listagem suspensa

• categoria errada na listagem

• listagens duplicadas

• informacao inconsistente online

• morada antiga ainda aparece

• avaliacoes sem resposta

• atributos em falta no perfil

12.15c Sector and situational variants

• hotel AI visibility

• clinic local search ranking

• salon Google Business Profile

• gym local SEO results

• retail store map ranking

• bar and nightlife local search

• cafe discovery searches

• bakery local visibility

• takeaway ordering visibility

• multi location restaurant group

• franchise local search

• single location independent business

• tourist facing business search

• expat facing business search

• seasonal business visibility

• high competition city local SEO

• low competition town local SEO

• new business no reviews

• established business stale profile

• rebranded business listing update

• relocated business listing

• visibilidade para hoteis

• SEO local para clinicas

• perfil Google para cabeleireiros

• ranking local para lojas

• negocios voltados para turistas

• negocio novo sem avaliacoes

• negocio que mudou de morada

12.15d Actions, features and adjacent surfaces

• Google Maps directions

• click to call listing

• reserve a table Google

• order online Google

• Google Maps filters

• open now filter

• price filter Maps

• rating filter Maps

• accessibility filter

• outdoor seating attribute

• accepts reservations attribute

• takeaway attribute

• delivery attribute

• vegetarian options attribute

• halal options attribute

• wheelchair accessible entrance

• parking availability attribute

• Wi-Fi attribute

• payment methods attribute

• live busy times

• popular times data

• photo view counts

• menu link Google

• website link UTM tagging

• Google posts weekly

• questions and answers seeding

• special hours holidays

• temporary closure handling

• business description 750 characters

• service area definition

• Apple Business Connect listing

• Bing Places listing

• TripAdvisor profile

• Trustpilot profile

• OpenTable listing

• delivery platform listings

• direcoes no Google Maps

• botao de chamada

• reservar mesa no Google

• filtros do Google Maps

• filtro aberto agora

• atributo esplanada

• atributo acessivel

• opcoes vegetarianas

• horarios especiais feriados

• descricao da empresa

• publicacoes semanais

• perguntas e respostas

• perfil no TripAdvisor

• listagem Apple Business Connect

12.15e Strategy and programme terms

• local SEO programme

• profile management retainer

• monthly optimisation cadence

• quarterly profile audit

• competitor monitoring local

• review workflow setup

• staff training review requests

• photography refresh schedule

• content calendar for posts

• multi surface visibility strategy

• entity data governance

• listing hygiene programme

• citation cleanup project

• ninety day local plan

• twelve month visibility roadmap

• baseline measurement setup

• reporting cadence agreement

• deliverable verification

• change log requirement

• programa de SEO local

• avenca de gestao de perfil

• auditoria trimestral do perfil

• monitorizacao de concorrentes

• fluxo de pedidos de avaliacao

• calendario de publicacoes

• plano de noventa dias

• roteiro de visibilidade

• registo de alteracoes

12.15f Comparison and evaluation terms

• SEO versus GEO versus AIO

• local pack versus AI answer

• Google versus ChatGPT recommendations

• Maps versus Search discovery

• mobile versus desktop local search

• reviews versus category relevance

• volume versus recency reviews

• profile optimisation versus website work

• organic versus paid local visibility

• in house versus agency local SEO

• DIY profile management

• free versus paid local tools

• single location versus multi location

• branded versus category demand

• snapshot versus grid measurement

• SEO versus GEO versus AIO comparado

• pacote local versus resposta de IA

• Google versus ChatGPT recomendacoes

• Maps versus Search descoberta

• movel versus computador

• avaliacoes versus relevancia de categoria

• gestao propria versus agencia

12.16 How to deploy this bank

Case study pages serve a different function

Almost nothing here is top-of-funnel. These terms are searched by people evaluating a supplier or diagnosing a problem they already have. Treat this page as a conversion asset supported by the educational guides, not as a traffic play in its own right.

Cluster 12.13 is where the demand sits

“Why does my competitor rank higher with fewer reviews” and “how do I get recommended by ChatGPT” are real questions with real volume and almost no good answers published. This case study answers both with evidence, which is unusual.

Link this page from the educational guides, not the reverse

The What Is Local SEO and What Is Off-Page SEO guides earn the traffic. This page converts it. Internal links should flow toward the proof, not away from it.

Validate locale before briefing

Portuguese-language volume for these terms skews Brazilian; Chinese-language terms need separate research entirely. Set tools to the market you are actually targeting.

Total terms in this bank: 555

13. Recommended Schema Markup for This Page

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Case Study: #1 on Google Search, AI Mode and ChatGPT",
  "description": "How Uma Marisqueria Macau became the top seafood recommendation across three search systems, with the Google Business Profile data behind it.",
  "author": {"@type": "Organization", "name": "Umarella", "url": "https://umarella.org"},
  "publisher": {"@type": "Organization", "name": "Umarella"},
  "dateModified": "2026-08-14",
  "about": [
    {"@type": "Thing", "name": "Local search engine optimization"},
    {"@type": "Thing", "name": "Generative engine optimization"}
  ]
}

Use Article for the case study page itself. Do not mark this page up with Restaurant or AggregateRating schema the rating belongs to the client’s own listing, not to your case study page, and misapplied review markup is a structured data violation.

14. The Takeaway

Visibility is no longer won on a single platform. The same customer, asking the same question, now has three ways to ask it and a business optimised only for the first can win the old game while quietly losing the new one.

What this case demonstrates is that the three do not require three separate strategies. They require one job done properly: accurate categories, complete fields, genuine review velocity, and information about the business that is consistent and machine-readable wherever it appears. Google’s local algorithm reads that. Google’s AI Mode reads it. ChatGPT, drawing on an entirely separate corpus, reads it too.

The evidence that this was the deciding factor rather than luck sits in the competitive context. A listing with 265 reviews placed above one with roughly 2,000. A perfect five-star average from a single review placed third. Whatever these systems are weighing, it is not the metric most business owners assume.

And the discovery data underneath makes the commercial case on its own: roughly seventeen non-branded searches for every branded one. Those are people who did not know this restaurant existed and found it anyway.

14.1 Where to start on your own listing

1. Check your primary category. Thirty seconds, free, and the single highest-leverage field there is.

2. Split your searches into branded and non-branded. If most are branded, you are a lookup destination rather than a discovery engine.

3. Ask the three engines your category question. Google AI Mode, ChatGPT, and one other. Record whether you appear.

Want the same audit on your business? Request a free Google Business Profile and AI visibility review umarella.org

Reservation at Uma Marisqueira Macau - 正宗葡式海鮮飯Reserve your spot at Uma Marisqueira Macau - 正宗葡式海鮮飯. We look forward to seeing you!umarella.org