--- title: "SEO & GEO Process in 2026-2027 : Strategy to Execution| Justin Hà" description: "A step-by-step SEO and GEO process: from business strategy and site structure to technical setup, content clusters, and AI citation tracking." url: https://justinha.info.vn/geo/seo-geo-process/ --- # SEO & GEO Process in 2026-2027 : Strategy to Execution| Justin Hà # The Complete SEO and GEO Process: From Strategy to Technical *Execution* Written by [Justin Hà](https://justinha.info.vn/about/) · August 18, 2026 · On This Page ![The Complete SEO and GEO Process: From Strategy to Technical Execution](https://justinha.info.vn/wp-content/uploads/2026/08/seo-geo-process.jpg) Take this article with you Copy as Markdown Copy link Ask an AI to summarise it Key Takeaways - 1. Understand the business, the user, the competitors, and the current brand baseline. - 2. Structure the site into Business Info, Service, and Information layers with consistent URLs. - 3. Fix the technical layer: sitemap, robots.txt, HTML-rendered content, and schema markup. - 4. Build content clusters of 20 to 30 topics per service, sorted by funnel stage, targeted narrow to broad around the real USP, linked correctly, and written answer-first. - 5. Build dedicated GEO signals: entity presence, llms.txt, extraction-ready content, platform-specific source research, and becoming a citable source. - 6. Test a fixed prompt set regularly to see how AI responds and which sources get cited. - 7. Keep the brand consistent across owned and external channels. - 8. Track SEO and GEO KPIs separately, using GSC, GA4, GTM, Clarity, and a dedicated GEO tool like WorkDuo. - 9. Audit, measure, and adjust on a fixed schedule. Over 5 years, I have completed more than 30 SEO and GEO projects across a wide range of industries and markets. More than 10 of these were dedicated GEO projects in hospitality, covering hotels, restaurants, tour agencies, property management, and the luxury travel segment, such as the work behind the [Tsar Carpets](https://justinha.info.vn/case-study/tsar-carpets/) and [Lady and the Hair](https://justinha.info.vn/case-study/lady-and-the-hair/) case studies, alongside [ecommerce brands](https://justinha.info.vn/geo-ai-branding/agentic-ecommerce/) outside hospitality. The work spanned global markets including the US, UK, Australia, Thailand, and Indonesia, as well as the domestic Vietnam market. Running that many projects across such different business types and regions keeps surfacing the same pattern. The process that actually works combines marketing thinking, brand thinking, business thinking, and technical skill together, not four separate departments working in isolation. That repeated pattern is what this article distills into one process. A complete SEO/GEO process has six stages: understand the business and audience, build the right website structure, fix the technical layer, produce content in organized clusters, build dedicated GEO signals, then keep the brand consistent and measure results. This article walks through the full process, from the strategic thinking behind each step to the exact technique used to execute it. ## Stage 1: Understand Before You Build Most SEO and GEO mistakes happen because this stage gets skipped. Before touching keywords, site structure, or code, get clear on four things: the business goal, the user’s real question, the direct competitors, and what AI tools already say about the brand. These answers shape every technical and content decision that follows. ### Step 1: Understand the Business Itself You cannot optimize what you don’t understand. Get clear on the product, the customer, and the current priority before any keyword research begins. **Key questions to answer:** - What does the business actually sell, and how does it make money? - Who is the ideal customer, and what do they care about most? - What is the single most important goal right now: awareness, leads, or sales? **How to do it:** Talk directly to the business owner or sales team. Read the existing About page, service pages, and any customer data available. Write down the business goal for the next 3 to 6 months. A business cannot chase every goal at once, so the SEO/GEO plan should support one clear priority per phase. ### Step 2: Understand the User’s Needs and Pain Points Once the business side is clear, flip the view to the customer side. This step finds the real problems people are trying to solve. **Key questions to answer:** - What problem is the customer solving before they even think about this product? - What words do they use to describe that problem, both in Google and in AI chat tools? - What stops them from deciding: price, trust, or missing information? **How to do it:** Review customer support tickets, sales call notes, and reviews. Check Reddit, Quora, and industry forums to see how people naturally phrase their questions. These real questions become the foundation for both keyword targeting and content topics in Stage 3. ### Step 3: Research Direct Competitors Not every website ranking in Google is a real competitor. This step finds the businesses that actually compete for the same customer. **Key questions to answer:** - Which businesses sell the same service, to the same audience, with the same goal? - What are they doing better: content depth, site structure, or brand trust? - Where are they weak or missing something entirely? **How to do it:** Search the core service keywords and shortlist 3 to 5 real competitors, not news sites, affiliates, or general blogs. Review their site structure, top pages, and content clusters using Ahrefs, Semrush, or manual search. Turn the findings into a simple gap checklist showing what they have that this site does not. ### Step 4: Run a Brand Audit. What Is AI Already Saying? Before building anything new, check the starting point. AI tools already have an opinion about the brand, even if no one has actively managed it. Our own audit of [how ChatGPT picks a brand before it even searches](https://justinha.info.vn/geo/chatgpt-decides-who-wins-before-it-searches/) is a real example of what this baseline check can uncover. **Key questions to answer:** - What does ChatGPT, Perplexity, or Google AI Overview currently say about this brand? - Is that information accurate, outdated, or missing? - How consistent is the brand’s information across owned channels (website, social) and external mentions (press, reviews, backlinks, forums)? **How to do it:** Ask an AI tool directly what it knows about the brand name. Compare the answer against reality. Separately, review owned channels for accuracy and search for external mentions to spot outdated or conflicting details. This audit becomes the baseline for measuring progress later, in Stage 6. ## Stage 2: Build the Website Foundation With the strategy clear, the next stage builds the structure and technical setup that both users and crawlers depend on. A website needs a clean skeleton before it needs more content. ### Step 5: Structure the Website Into Three Layers Every page on the site should belong to one of three layers. Mixing them confuses both users and search engines. Layer Example URL Purpose Business Info /about/, /contact/, /policy/ Prove the business is real and trustworthy Service /services/[service-name]/ Explain exactly what is sold Information /guides/[service-name]/[topic]/ Answer questions and support the related service Keep this URL pattern consistent across the whole site. Consistency helps both search engines and AI crawlers understand which articles belong to which service, and it makes internal linking easier to manage at scale. Keep the Business Info and Service layers fully transparent and up to date, since AI tools rely heavily on this information to describe the business correctly. **Anatomy of a Service page**, in the pattern used for pages like [the hotel and resort service page](https://justinha.info.vn/geo-ai-branding/hotel-resort/): 1. H1 with the exact service name. 2. Opening answer (40 to 60 words) stating what the service does and who it’s for, with no introduction fluff. 3. What’s included: a bullet list of deliverables. 4. Process: a short numbered list of how the service works. 5. Proof: results, data, or case studies if available. 6. FAQ block with 3 to 5 real pre-purchase questions. 7. Links to 3 to 5 relevant Information articles from the same cluster. **Anatomy of an Information page:** 1. H1 phrased as the actual question, for example “How Much Does a GEO Audit Cost?” 2. Direct answer in the first 30 to 50 words. 3. Supporting explanation broken into H2/H3 sections, one sub-idea each. 4. A real example or data point. 5. One clear link back to the related Service page. ### Step 6: Fix the Technical Layer A well-structured site still needs to be technically readable by crawlers, including AI crawlers, which often [behave differently from Googlebot](https://justinha.info.vn/geo/how-ai-bots-reach-your-website/). Sites that need this layer fixed without a full retainer can use the [Technical GEO Package](https://justinha.info.vn/geo-ai-branding/technical-package/) instead. **Key questions to answer:** - Does the main content load immediately in the raw HTML, or only after scripts run? - Does the sitemap accurately reflect the site’s real, indexable pages? - Does robots.txt correctly allow the crawlers this site wants, including AI bots? - Does structured data clearly label what each page contains? **Checking if content renders without JavaScript:** Right-click the page, choose View Page Source, and search for the main heading text. If it isn’t there, the content is being injected after load by JavaScript, and some AI crawlers may see a blank page. Move critical content to render directly in the HTML. Our [log file guide](https://justinha.info.vn/geo/what-is-seo-log-file/) covers how to confirm which crawlers actually reach the page at all. > **Don’t check this by hand.** The [AI Log File Analyzer](https://justinha.info.vn/ai-log-file-analyzer-tool/) reads your server logs and shows exactly which AI bots crawled which pages, and how often, so you don’t have to grep logs manually. **XML sitemap setup:** For sites with more than a few hundred pages, split the sitemap by content type instead of using one large file. - `/sitemap-pages.xml` for Business Info pages - `/sitemap-services.xml` for Service pages - `/sitemap-articles.xml` for Information and cluster articles - `/sitemap-index.xml` as the master file linking to all of the above Only include canonical URLs that return a 200 status code, exclude noindex and duplicate parameter URLs, and make sure the sitemap updates automatically when pages are published or removed. Submit the sitemap index URL in Google Search Console. The [AI Sitemap Analyzer](https://justinha.info.vn/ai-sitemap-analyzer-tool/) checks all of this in one pass instead of opening every sitemap file by hand. **Robots.txt setup:** A basic, safe configuration looks like this. ``` User-agent: * Disallow: /wp-admin/ Allow: /wp-admin/admin-ajax.php Disallow: /cart/ Disallow: /checkout/ Disallow: /*?s= Sitemap: https://example.com/sitemap-index.xml ``` For GEO, decide explicitly whether to allow [AI crawlers such as GPTBot](https://justinha.info.vn/geo/openai-bots/) (OpenAI), ClaudeBot (Anthropic), PerplexityBot (Perplexity), and Google-Extended (controls use in Google’s AI features). To allow one: ``` User-agent: GPTBot Allow: / ``` Replace `Allow: /` with `Disallow: /` to block instead. Blocking protects content from AI training use but may reduce the chance of being cited in AI answers. > **This is a business decision, not only a technical one.** Weigh training-data protection against citation visibility before blocking any AI crawler outright, and check your [crawl budget data](https://justinha.info.vn/geo/ai-bot-crawl-budget-log-file/) first to see how much these bots are actually requesting. **Schema markup (JSON-LD):** Add structured data in the page head so search engines and AI systems know exactly what each page is. Organization schema for the homepage: ``` { "@context": "https://schema.org", "@type": "Organization", "name": "Business Name", "url": "https://example.com", "logo": "https://example.com/logo.png", "address": { "@type": "PostalAddress", "streetAddress": "Street", "addressLocality": "City", "addressCountry": "VN" } } ``` Service schema for Service pages: ``` { "@context": "https://schema.org", "@type": "Service", "name": "Service Name", "provider": { "@type": "Organization", "name": "Business Name" }, "areaServed": "Vietnam", "description": "Short description of the service." } ``` FAQPage schema for any page with a visible FAQ section: ``` { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "Question text here", "acceptedAnswer": { "@type": "Answer", "text": "Answer text here" } } ] } ``` Article schema for Information and cluster pages: ``` { "@context": "https://schema.org", "@type": "Article", "headline": "Article title", "author": { "@type": "Person", "name": "Author Name" }, "datePublished": "2026-01-01", "dateModified": "2026-08-01" } ``` Most WordPress SEO plugins generate basic Article schema automatically, but FAQ and Service schema usually need to be added manually or through a dedicated schema plugin, since they depend on each page’s actual content. ## Stage 3: Produce Content in Organized Clusters With the structure and technical layer in place, content production can begin. A cluster is one Service page (the pillar) plus 20 to 30 supporting Information articles built around it, enough to cover the topic from every angle a customer might search for. Teams that need this cluster written rather than built in-house can use the [Content Package](https://justinha.info.vn/geo-ai-branding/content-package/) service to produce it. ### Step 7: Build Keyword Clusters and Write Answer-First Content **Key questions to answer:** - What are the 20 to 30 real questions customers ask about this service? - Where does each question fall in the buyer journey: awareness, consideration, or decision? - Does every article answer its main question in the first sentence? **Finding the 20 to 30 topics:** Pull ideas from five sources, then combine and deduplicate into one list. 1. **Customer questions.** From support tickets, sales calls, and messages. 2. **Search data.** From Google Keyword Planner, Ahrefs, or Semrush, including “People Also Ask” results. 3. **Community platforms.** Reddit, Quora, and industry forums. 4. **AI query mining.** Ask ChatGPT or Perplexity what people typically want to know about this type of service. 5. **Competitor gap.** Topics competitors cover that this site doesn’t have yet. **Sorting topics by funnel stage:** Stage Example Topics Goal Awareness (about 40%) “What is GEO?”, “Why does AI search matter?” Explain the basic concept Consideration (about 35%) “GEO vs traditional SEO”, “How to choose a GEO consultant” Help the reader compare options Decision (about 25%) “GEO audit pricing”, “What happens in a GEO audit” Push toward a decision Decision-stage articles should link most directly and most often back to the Service page. An “awareness” article, for example, is exactly what [our own “what is GEO” guide](https://justinha.info.vn/geo/what-is-generative-engine-optimization/) is built to be. **Internal linking rules within a cluster:** - Every supporting article links back to its pillar (Service) page at least once. - The pillar page links out to its 5 to 8 most important supporting articles, not all 20 to 30. - Articles within the same funnel stage may link to each other when genuinely relevant. - Vary the anchor text across links instead of repeating the same phrase. - Limit each content section to one internal link, and avoid stacking several links in one paragraph. **Writing technique for every page:** - Answer the main question in the first 1 to 2 sentences, before any introduction. - Keep paragraphs to 60 to 100 words and sentences under 20 words. - One idea per paragraph. Start a new paragraph for each new idea. - Write headings as real questions where natural, and answer them directly underneath. - Use numbers instead of vague claims, for example “reduces cost by 35%” instead of “very effective.” - Cut filler phrases such as “it is important to note that” or “in today’s world.” ### Step 8: Target by Service and USP, Not by Generic Category A generic keyword attracts a broad, mixed audience with mixed intent. Targeting the business’s actual unique selling point attracts the exact customer that service was built for, and it is far easier to rank and get cited for a specific angle than a crowded generic term. **Key questions to answer:** - What is this service’s real USP, not just its category? - Who benefits most specifically from that USP? - Which narrow version of the broad keyword reflects that USP? **How to do it:** Instead of targeting a broad term like “best hotel in Da Nang,” identify the actual strength of the business, for example a hotel with a kids’ club and family suites, and target “best hotel in Da Nang for families with 2 kids” first. Build the cluster around that specific audience: room setups for families, activities for children, family-friendly dining, safety features. Once that narrow cluster ranks and gets cited consistently, expand outward to slightly broader terms, and eventually compete on the generic term with a much stronger topical base behind it. This narrow-to-broad approach applies to every service with a real USP, not just hospitality, as shown in the [Cucina Luca case study](https://justinha.info.vn/case-study/cucina-luca/). ## Stage 4: Build Dedicated GEO Signals Traditional SEO and clean content structure already help AI tools understand a site. This stage goes further, targeting the specific signals AI systems use to decide what to trust and what to cite. ### Step 9: Build the Brand’s Entity Presence AI models learn about brands from many sources across the internet, not just the brand’s own website. A strong entity presence makes it easier for AI to recognize and correctly describe the business. **Key questions to answer:** - Does the brand have a presence on data sources AI models commonly draw from, such as Wikidata, Crunchbase, or industry directories? - Does the business appear correctly on Google Business Profile and major review platforms? - Is the brand description worded the same way everywhere it appears? **How to do it:** Create or claim listings on Wikidata, Crunchbase, LinkedIn Company pages, and relevant industry directories. Use the exact same business name, description, and category everywhere. Keep Google Business Profile fully filled out and active, since it feeds both Google’s Knowledge Graph and several AI tools, the same local-business layer covered in [our analysis of how ChatGPT pulls in Maps data](https://justinha.info.vn/geo/chatgpt-pulls-in-google-maps-data-of-answers/). This entity work matters most for high-consideration purchases like [real estate](https://justinha.info.vn/geo-ai-branding/real-estate/) and [property management](https://justinha.info.vn/geo-ai-branding/property-management/), where a buyer researches for weeks before ever making contact. > **Check which external sites are actually vouching for the brand.** The [AI Backlink Audit Tool](https://justinha.info.vn/ai-backlink-audit-tool/) shows which backlinks and mentions AI models are likely to weigh, instead of guessing from a generic backlink report. ### Step 10: Set Up llms.txt [llms.txt works like robots.txt](https://justinha.info.vn/geo/what-is-llms-txt-llms-full-txt-cats-txt/), but it is aimed specifically at AI agents. It tells them which content to prioritize and how it may be used. **How to do it:** Create a plain text file at the root of the domain, for example `https://example.com/llms.txt`, listing the site’s most important pages with a short description of each. ``` # example.com llms.txt ## Business Info - About: https://example.com/about/ - Contact: https://example.com/contact/ ## Services - GEO Audit: https://example.com/services/geo-audit/ ## Guides - What is GEO: https://example.com/guides/geo-audit/what-is-geo/ ``` This standard is still new and not every AI platform reads it yet, but adoption is growing, and setting it up early costs very little effort. ### Step 11: Write Extraction-Ready Q&A Content AI tools favor content that can be lifted out and used as a direct answer, without needing extra editing. This step makes each answer self-contained. **Key questions to answer:** - Can this paragraph be copied out and still make complete sense on its own? - Does the article clearly define its key term the first time it appears? - Does a longer article end with a short summary an AI model could quote directly? **How to do it:** For definitions, start the relevant section with a “Definition:” label followed by one clear sentence. For longer guides, add a short “Summary:” section near the end that restates the main points in plain language. Keep each supporting paragraph between 60 and 100 words so it can stand alone as a quotable block. ### Step 12: Understand How Each AI Platform Chooses Its Sources Each AI platform pulls citations from a different mix of sources, based on how it retrieves and ranks information. Treating every platform the same wastes effort, since the source that gets an answer cited on one platform may be ignored by another. **Key questions to answer:** - Which source types does each target AI platform tend to cite most often? - Is the brand present and active on those specific source types? - Which platform matters most for this business’s actual customers? AI Platform Commonly Cited Sources What This Means for Optimization Google AI Overview Own indexed pages, Facebook Groups, YouTube, established forums and media Strong on-page schema plus a real presence in relevant Facebook Groups can help ChatGPT (with browsing) Reddit threads, Wikipedia, review sites, established publications Genuine participation on Reddit and a solid Wikipedia or Wikidata presence matter Perplexity Recent news sites, structured data-rich pages, forums, the site’s own content Fresh, frequently updated, well-structured pages tend to perform better here These patterns shift over time as each platform updates its retrieval methods, as documented in our own before/after tracking of [GPT-5.6’s citation behavior change](https://justinha.info.vn/geo/gpt-5-6-killed-listicle-citations/), so treat this table as a starting point, not a fixed rule. Step 14 below explains how to verify current behavior directly. ### Step 13: Become a Source AI Actually Cites Yes, a business can influence what AI tools say by becoming one of the sources those tools already trust and pull from. This works by showing up, with real value, on the exact source types identified in Step 12. **Key questions to answer:** - Where can this brand contribute real, non-promotional value on the platforms its target AI tools cite most? - Does the business qualify for a Wikipedia entry, and if not yet, is a Wikidata entry realistic? - Is there original data or a unique angle worth publishing that other sources don’t have? **How to do it:** If Reddit matters for the target AI platform, participate genuinely in relevant subreddits under a real account, answering questions helpfully rather than posting links. If Facebook Groups matter, join and contribute to active groups in the niche rather than creating a group just to post promotions. Publish at least one piece of original data or a small study that other sites don’t have, since unique data gets cited more often than repeated general advice. None of this replaces the website itself, but it adds outside sources that AI models can find and trust alongside it. > **Being crawled is not the same as being cited.** A site can be fully crawled by every AI bot and still never appear in an answer. See [why that gap happens and how to fix it](https://justinha.info.vn/geo/ai-crawled-not-cited-fix/) before assuming the technical layer is the problem. ### Step 14: Build a Prompt Set and Test How AI Responds Without direct testing, there is no way to know whether any of this GEO work is actually working. This step turns testing into a repeatable process. **Key questions to answer:** - Which prompts actually matter to this business, covering branded, unbranded, comparison, and “best” style questions? - Does the brand appear in the answer, and if not, which competitor does? - Which source does the AI cite, and is the brand’s own website ever among them? **How to do it:** 1. **Build the prompt list.** 20 to 50 prompts a real customer might type into an AI tool: branded questions, unbranded category questions, “best X” questions, “X vs Y” comparisons, and alternatives or pricing questions. 2. **Run it across platforms.** The same prompt set across the AI platforms that matter most to this audience, for example ChatGPT, Perplexity, and Google AI Overview. 3. **Record three things per prompt.** Whether the brand appears, exactly which source is cited, and whether that source is the brand’s own website or an external one. 4. **Repeat on a fixed schedule.** Weekly or monthly, tracking whether visibility improves, stays flat, or drops. 5. **Act on the pattern.** If a competitor’s page is consistently cited instead, study why. If Reddit or a Facebook Group is cited instead of any brand website, that confirms where Step 13 effort should go next. This can be done manually with a spreadsheet for a small prompt set, but it becomes time consuming at scale. Stage 6 covers how to automate it. ## Stage 5: Build a Consistent Brand ### Step 15: Make the Brand Consistent Everywhere Both Google and AI tools build trust in a brand by checking whether its information matches across many sources. Inconsistency reads as a weak or unreliable signal. **Key questions to answer:** - Does the brand name, address, and service description match across the website, social profiles, and business directories? - Is the brand mentioned on any credible external sites, publications, or forums? - Are there outdated or conflicting details anywhere online? **How to do it:** Audit the brand’s name, address, contact details, and service descriptions across the website, Google Business Profile, social media, and directories. Fix mismatches. Then focus on earning a small number of high-quality mentions from relevant industry sites, local press, or expert comments, rather than spreading thin across hundreds of low-value profiles, the same consistency that carried [the Holiday Inn Saigon Airport result](https://justinha.info.vn/case-study/holiday-inn-saigon-airport/). ## Stage 6: Measure and Repeat ### Step 16: Track the Right KPIs for SEO and GEO Separately SEO and GEO are measured differently, and mixing their KPIs together hides what is actually happening. A page can lose ranking position while gaining AI citations, or the other way around, so both need their own framework. Dimension SEO KPI Framework GEO KPI Framework Primary signal Ranking position in search results Mention or citation rate inside AI-generated answers Traffic metric Organic sessions and click-through rate Referral traffic from AI platforms and share of voice against competitors Success measure Keyword rankings, backlinks earned Mention rate, sentiment of the mention, and accuracy of the brand description Typical timeframe Weeks to months to see movement Can shift within days as models update retrieval results Attribution A visible click leading to the site Often zero click, meaning brand influence without a site visit Set separate goals for each. For example, an SEO goal might be reaching page one for 10 priority keywords in 6 months, while a GEO goal might be appearing in 60% of the tracked prompt set within the same period. ### Step 17: Set Up the Measurement Stack **Key questions to answer:** - Is search performance being tracked at the query level, not just at the page level? - Can traffic coming specifically from AI platforms be identified separately from normal organic traffic? - Is there a dedicated way to track GEO metrics like mention rate and share of voice, without manually testing every prompt by hand? - **Google Search Console (GSC):** Use it to track indexing status, impressions, and clicks by query. This remains the primary source for confirming whether pages are indexed and how they perform in traditional search. - **Google Analytics 4 (GA4):** Set up a segment or exploration that filters referral traffic from AI platform domains, such as chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com, to see how many visitors actually arrive from AI tools and how they behave once on the site. - **Google Tag Manager (GTM):** Use it to set up custom event tracking, such as tagging conversions that originate from AI-referred sessions, so those visitors can be measured against normal organic visitors. - **Microsoft Clarity:** A free tool for session recordings and heatmaps. Use it to compare on-site behavior between AI-referred visitors and traditional search visitors, since the two groups often behave differently. - **GEO Audit Tool:** A free, dedicated GEO tracking tool built for exactly this job. It automates the prompt testing described in Step 14, checking citation rate, site: exposure, and entity presence without testing every prompt manually. - **GEO AI SEO Analytic Tool:** Complements the audit tool with ongoing analytics once a baseline is established, useful for the monthly or quarterly cycle in Step 18. Together, GSC and GA4 cover traditional SEO performance, GTM connects tracking to real business events, Clarity shows how AI-referred visitors actually behave, and the GEO Audit Tool covers the GEO-specific tracking that the other tools were not built for. > **For ongoing, multi-competitor tracking**, ChatGPT Search Tracker (an internal tool, not public yet) automates fan-out capture, site: filter detection, and citation benchmarking across models over time, the same tooling behind our [gpt-5-6 vs gpt-5-6-mini analysis](https://justinha.info.vn/geo/gpt-5-6-vs-gpt-5-6-mini/). ### Step 18: Audit, Measure, and Adjust on a Regular Cycle SEO and GEO are not one-time projects. This final step closes the loop and turns the process into a repeatable system. **Key questions to answer:** - Has traffic and ranking improved since the last review? - What is AI now saying about the brand compared to the original audit in Step 4? - What still needs fixing: content gaps, trust signals, or technical issues? **How to do it:** Repeat the brand audit from Step 4 and the prompt testing from Step 14 every month or quarter, and compare results against the KPI framework from Step 16. Also review Service pages and top-performing Information pages every 90 days for accuracy and freshness, and re-check the sitemap and robots.txt after any major site redesign. Use the findings to decide what to fix next, then repeat the cycle. This entire 18-step process can be run in-house once a team understands the framework. For anyone who wants to learn and run it themselves rather than outsource it, the [GEO Course](https://justinha.info.vn/geo-ai-branding/geo-course/) walks through this exact process step by step. ## Run This Process on Your Own Site Every stage above, technical audit, sitemap health, backlink signals, AI crawl behavior, and prompt tracking, maps to one of our free tools. Start with whichever gap this article surfaced first for your own site. Stage in this article Tool that covers it Step 6: technical layer, log files [AI Log File Analyzer](https://justinha.info.vn/ai-log-file-analyzer-tool/) Step 6: sitemap health [AI Sitemap Analyzer](https://justinha.info.vn/ai-sitemap-analyzer-tool/) Step 9: entity and backlink signals [AI Backlink Audit Tool](https://justinha.info.vn/ai-backlink-audit-tool/) Step 14 & 17: prompt testing, citation tracking [GEO Audit Tool](https://justinha.info.vn/geo-audit-tool/) If the gap is bigger than a single tool, our [GEO and AI visibility service](https://justinha.info.vn/geo-ai-branding/) runs this entire 18-step process end to end, with dedicated tracks for [restaurants](https://justinha.info.vn/geo-ai-branding/restaurant/) and [tour agencies](https://justinha.info.vn/geo-ai-branding/tour-agency/). [Get in touch](https://justinha.info.vn/contact/) to scope which stage to start with. ## Frequently Asked Questions **Is GEO replacing SEO?** No. GEO builds on solid SEO fundamentals such as clear structure, fast loading, and quality content. It adds extra attention to how AI tools read and cite that same content. **Can a business actually influence what AI says about it?** Yes, indirectly. A business cannot edit an AI model’s answer directly, but it can influence the sources that model relies on, by improving its own site, becoming active on the platforms that model cites, and earning a presence on the specific data sources listed in Step 12. **How long does this process take to show results?** Most businesses following this process consistently start seeing early signals within 3 to 6 months, depending on competition and starting point. AI visibility can shift faster than traditional rankings, but there is no guaranteed timeline in either SEO or GEO. **Do I need to redo the whole process for every new service?** No. Stage 1 is mostly done once at the business level. Stage 3 and Stage 4 repeat for each new service, while Stage 6 runs continuously across the whole site. --- ![Justin Hà](https://justinha.info.vn/wp-content/uploads/2025/11/justin-ha-seo-specialist-hcm.jpg) About the Author **Justin Hà** Senior Global SEO & GEO Specialist · Founder of Luminal Ho Chi Minh City, Vietnam Justin Hà is a Senior Global SEO & GEO Specialist with 5+ years of experience driving organic growth across 9 international markets. He is among Vietnam's first practitioners of Generative Engine Optimization (GEO) - optimizing brand visibility inside AI-generated answers from ChatGPT, Gemini, Perplexity, and other platforms. In 2026, he founded [Luminal](https://www.linkedin.com/company/luminal-geo/), a GEO agency helping brands get discovered, cited, and trusted inside AI-driven search. With 2+ years in GEO and 100+ projects led or consulted, Justin helps brands adapt their content strategies to the rapidly evolving AI search landscape. [Vietnam's First GEO Practitioner](https://justinha.info.vn/about/) 5+ Years in SEO 2+ Years in GEO 100+ Projects 10+ Active GEO [ ](https://www.linkedin.com/in/justinhaseo/) [ ](https://www.linkedin.com/company/luminal-geo/) [ ](https://www.facebook.com/justinhaseo) [ ](https://x.com/justinhaseo) [ ](https://www.instagram.com/justinhaseo/) ## Related Articles [ ![GPT-5.6 Listicle Citations: What Changed, What Didn’t](https://justinha.info.vn/wp-content/uploads/2026/09/listicle-format-citation-chatgpt-768x432.jpg) GEO GPT-5.6 Listicle Citations: What Changed, What Didn’t September 19, 2026 ](https://justinha.info.vn/geo/gpt-5-6-killed-listicle-citations/) [ ![ChatGPT Pulls In Google Maps Data for 75% of Answers, Here’s What That Means](https://justinha.info.vn/wp-content/uploads/2026/09/local-map-result-in-chatgpt-768x427.jpg) GEO ChatGPT Pulls In Google Maps Data for 75% of Answers, Here’s What That Means September 18, 2026 ](https://justinha.info.vn/geo/chatgpt-pulls-in-google-maps-data-of-answers/) [ ![ChatGPT’s Two Models Search Very Differently, But Cite the Same Winner](https://justinha.info.vn/wp-content/uploads/2026/09/site-search-in-chatgpt-768x410.png) GEO ChatGPT’s Two Models Search Very Differently, But Cite the Same Winner September 18, 2026 ](https://justinha.info.vn/geo/gpt-5-6-vs-gpt-5-6-mini/)