Strategy and procurement
GEO, SEO and AEO: What to Commission and What to Ask First
A practical way to assess the scope, deliverables and evidence in an AI visibility proposal for Chile and Latin America.
3 min readExploreResources
Working guides for teams in Chile and Latin America. We distinguish official documentation, research findings and SERVIGEO recommendations. Each source states its scope and review date.
Strategy and procurement
A practical way to assess the scope, deliverables and evidence in an AI visibility proposal for Chile and Latin America.
3 min readExploreMeasurement
Design a question panel, retain the answers and separate observations, official reports and commercial outcomes.
3 min readExploreImplementation
Prioritize verifiable barriers, business identity, sources and checks after publication.
3 min readExploreA reusable record of questions, sources and actions to keep each review traceable.
These definitions explain how to read a report. Proprietary metrics must publish their formulas and must never be presented as internal Google, OpenAI or Bing scores.
The result of documented checks on website access, structure and content.
It does not measure whether an AI already knows, mentions or recommends the business. A score requires visible criteria, weights and evidence.
Valid answers that name the brand, divided by valid answers in the panel, multiplied by 100. Each answer counts at most once for that brand.
It describes only the sample and its conditions. If there are no valid answers, report no data.
The percentage of valid answers in the panel that include at least one verifiable link to the domain under review.
A name without a link is a mention. This rate is neither Bing's official citation count nor a global share.
Verified, accurate commercial claims divided by all claims checked. Evidence is retained for every review.
Unverified claims are reported separately. They are not automatically classified as accurate or false.
Appearances of links to the website under the aggregation rules of Search Console's generative report.
Property-level and page-level totals may differ. They are not equivalent to clicks or commercial enquiries.
Visible references to content in answers from supported experiences during the selected period.
These are aggregated and sampled data, rather than an exhaustive answer log or visit count.
A grouped phrase related to retrieving cited content.
It is not necessarily a user's exact question.
The percentage of citations attributed to the website out of all citations for the same grounding query.
It does not show competitor domains, traffic share or content quality.
Sessions whose origin is identified using available signals and documented attribution rules.
It does not necessarily capture the customer's entire journey. An unknown origin must remain identified as unknown.
A contact that meets agreed criteria for need, coverage and capacity to serve, reviewed by the sales team.
Attribution requires its own evidence. It must not be inferred from a mention, impression or citation.
Reviewed: 2026-09-07 · These sources support specific statements; they do not imply a partnership with SERVIGEO.
Google retains SEO as the foundation of its generative experiences. Eligibility requires indexing, snippet eligibility and inclusion in Search's generative features; llms.txt and special markup offer no advantage in Google.
Meeting these requirements does not guarantee indexing or appearance. This guide describes Google Search; it does not define other assistants' rules.
Optimizing your website for generative AI features on Google Search ↗The report documents AI Overviews and AI Mode impressions by page, country, date and device. Google announces its global rollout from August 31, 2026; a site may not show data if activity is insufficient.
The reviewed documentation describes impressions, rather than a separate breakdown of AI clicks, CTR or positions. It excludes Search Labs experiments.
Generative AI performance report (Search) ↗Search Console allows site links and content to be included in or excluded from generative features in Search and Discover. Inclusion is the documented default setting.
The inclusion setting establishes eligibility, rather than promising traffic or a citation for any page. Reviewing it requires access to the relevant property.
Search generative AI control ↗OAI-SearchBot serves search; GPTBot crawls content that may be used for training. Their preferences are independent. ChatGPT-User handles certain user actions and does not determine inclusion in Search; robots.txt may not apply to those visits.
Allowing an agent does not guarantee a citation. Do not claim that blocking GPTBot prevents appearance in ChatGPT Search or that a ChatGPT-User visit proves indexing.
Overview of OpenAI Crawlers ↗Bing introduced data on citations, cited pages and grounding queries for Copilot, Bing summaries and selected partner integrations. The metrics describe activity within supported experiences.
This is not a census of all assistants. Citation counts do not describe a page's position, authority or importance within an answer.
Introducing AI Performance in Bing Webmaster Tools Public Preview ↗Bing added Intents, Topics, Citation Share and Compare in global preview. Citation Share shows the site's proportion of citations for a grounding query; the new views reveal context and changes over time.
It does not reveal competitor domains or represent traffic share or a quality score. Time comparisons reveal changes but do not establish their cause.
New AI Visibility Insights in Bing Webmaster Tools: Intents, Topics, Citation Share, Compare ↗The help page explains that citation and grounding-query data are aggregated and sampled. Queries group retrieval phrases; they do not show users' complete questions. The data supports trend analysis and exports.
A citation is not a click. A missing query may reflect low activity and does not establish a penalty; different report breakdowns may produce different totals.
AI Performance in Bing Webmaster Tools ↗The study evaluated content changes and reported improvements of up to 40 % in experimental visibility metrics, with results varying by domain. It used an experimental engine and a Perplexity evaluation with texts uploaded as files across 200 samples.
It does not establish an increase in sales or clicks, or a 40 % improvement for current Latin American clients. The experiment's metrics are not market shares; the authors acknowledge that methods must adapt as engines change.
GEO: Generative Engine Optimization ↗SERVIGEO / Your next step
Tell us about your business. Let’s define where to begin.