AI SEO Roadmap: Your First 90 Days (Week-by-Week Plan)

Before You Start: What You Need

If you are new to AI SEO, here is the short version: it is the practice of making your website visible to AI assistants like ChatGPT, Gemini, Perplexity, and Claude so they cite your content when answering user questions. Traditional Google rankings do not guarantee AI visibility — 88% of pages cited by AI models are not in Google's top 10.

This roadmap assumes you have a live website with at least 10-15 published pages, CMS access to edit content and upload files, a Google Analytics 4 account (or the ability to set one up), and access to your robots.txt file (either through your CMS or server).

You do not need coding skills, a large budget, or prior AI SEO experience. You do need consistent execution — plan for 5-12 hours per week over the next 12 weeks.

Your first task before starting week 1: Run a free AI visibility scan at AImetrico and write down your AI Score. This is your baseline. Every milestone in this roadmap measures improvement against this number.

Weeks 1-2: Baseline and Critical Fixes

Goal: Understand exactly where you stand and fix the issues that make you completely invisible to AI.

Time investment: 5-8 hours per week.

Week 1: Full Baseline Audit

Task 1: Check your current AI visibility across all platforms. Open ChatGPT, Gemini, Perplexity, and Claude. Ask each one questions that your ideal customer would ask — questions your website should be answering. Note which platforms mention you, which mention competitors instead, and which give inaccurate information about your brand. Use our multi-platform check guide on how to see if your site is visible in AI for the exact queries to test.

Task 2: Audit your robots.txt file. This is the single most common cause of AI invisibility. Open yoursite.com/robots.txt in a browser and check whether AI search bots — specifically OAI-SearchBot, PerplexityBot, and ChatGPT-User — are blocked. Many sites have blanket Disallow: / rules that prevent all AI crawlers from seeing any content. Our robots.txt for AI crawlers guide walks through every line you need.

Task 3: Audit your existing schema markup. Run your homepage and top 5 pages through Google's Rich Results Test. Document which schema types are present (if any), whether they validate without errors, and what is missing. You will fix these in weeks 3-4.

Task 4: Measure page speed. Run your top 10 pages through Google PageSpeed Insights. Record the First Contentful Paint (FCP) and Largest Contentful Paint (LCP) scores. Sites with FCP under 0.4 seconds are cited by ChatGPT 3x more often. Flag any page with FCP above 1.5 seconds as a priority fix.

Expected outcome by end of week 1: A document listing your AI Score baseline, robots.txt status, schema gaps, and page speed scores. This is your "before" snapshot.

Week 2: Critical Fixes

Task 5: Fix your robots.txt. If AI search bots are blocked, update your robots.txt immediately. Allow OAI-SearchBot, ChatGPT-User, PerplexityBot, Google-Extended (for Gemini), and ClaudeBot. You can choose to block training-only bots (GPTBot, CCBot) while keeping search bots open. This single change can make your entire site visible to AI within 1-2 weeks.

Task 6: Fix critical page speed issues. For any page with FCP above 1.5 seconds, address the biggest bottlenecks: compress images, enable browser caching, defer non-critical JavaScript, and eliminate render-blocking resources. You do not need to achieve perfect scores — getting FCP below 1.0 seconds is sufficient for AI crawlers.

Task 7: Fix broken or invalid existing schema. If your baseline audit found schema markup with errors, fix the validation issues now. Broken schema is worse than no schema — it confuses AI models about your content.

Task 8: Document your top 5 target pages. Identify the five pages on your site that matter most for AI visibility. These are typically your homepage, your primary product or service page, your about page, and your two best-performing content pieces. These pages will be the focus of your optimization in weeks 5-6.

Expected outcome by end of week 2: AI crawlers can access your site (robots.txt fixed), your pages load fast enough for AI retrieval, and you have a prioritized list of pages to optimize.

Weeks 3-4: Technical Foundation

Goal: Build the structured data and tracking infrastructure that AI models rely on to understand and cite your content.

Time investment: 6-8 hours per week.

Week 3: Schema Markup Implementation

Task 9: Implement Organization schema on your homepage. This tells AI models who you are as an entity — your name, logo, contact information, social profiles, and founding details. Organization schema is the single most important schema type for brand recognition in AI responses. Start with our JSON-LD basics for AI SEO guide if you have not worked with structured data before.

Task 10: Add Article schema to all blog posts and content pages. Every piece of content should have Article (or TechArticle, NewsArticle, or BlogPosting) schema with headline, author, datePublished, dateModified, and description. This is how AI models determine whether your content is current and trustworthy.

Task 11: Add FAQPage schema to pages with question-and-answer content. If a page answers common questions — even implicitly — add FAQ schema. Research shows that FAQ schema improves AI content interpretation from 16% to 54%. You can add FAQ sections to existing pages specifically for this purpose (you will do this at scale in weeks 5-6).

Task 12: Add author bios to all content pages. Create or update author bio sections with real names, credentials, job titles, and relevant expertise. Link author bios to LinkedIn profiles where possible. AI models use E-E-A-T signals heavily when deciding which sources to cite — and author expertise is one of the strongest signals. Add Person schema markup to each author bio.

Week 4: llms.txt and Tracking Setup

Task 13: Create and publish your llms.txt file. This is a relatively new standard — a plain text file at yoursite.com/llms.txt that tells AI models about your site, its purpose, its key pages, and how to interpret your content. Think of it as a robots.txt specifically for AI understanding rather than crawling. Follow our complete llms.txt specification guide for the exact format and recommended content.

Task 14: Set up GA4 AI referral traffic tracking. Configure Google Analytics 4 to identify and segment traffic from AI platforms: chatgpt.com, perplexity.ai, claude.ai, copilot.microsoft.com, and gemini.google.com. Create a custom channel group for "AI Referral" traffic so you can measure it separately from organic and direct. Our GA4 AI referral traffic setup guide has step-by-step instructions with screenshots.

Task 15: Validate all new schema markup. Run every page you modified through Google's Rich Results Test and Schema.org's validator. Fix any errors. Then test the pages manually in ChatGPT and Perplexity — ask questions that your pages should answer and see if the AI can now find and interpret your content.

Task 16: Create a monitoring spreadsheet. Set up a simple tracking sheet with columns for: date, AI Score, ChatGPT mentions (yes/no for top 5 queries), Gemini mentions, Perplexity mentions, and AI referral sessions from GA4. You will update this weekly from now through day 90.

Expected outcome by end of week 4: Your site has comprehensive schema markup (Organization, Article, FAQPage, Person), a live llms.txt file, GA4 tracking for AI referral traffic, and a monitoring system in place.

Milestone: Day 30 Checkpoint

Stop and measure. Run your AI visibility scan again and compare your new AI Score to your baseline from day 1.

What to expect at day 30:

• AI Score improvement: 10-15 points above baseline (primarily from technical fixes)
• robots.txt: All AI search bots unblocked and confirmed crawling your site
• Schema coverage: Organization schema on homepage, Article schema on all content pages, FAQ schema on at least 3 pages
• Page speed: All top 10 pages with FCP below 1.0 seconds
• Tracking: GA4 showing AI referral traffic (even if numbers are small)
• llms.txt: Published and accessible

If you are behind: The most common cause of slow progress at day 30 is robots.txt changes not propagating. Verify your changes are live (not cached) and check server logs for AI crawler activity. If AI bots are visiting but your score has not improved, your schema markup may have validation errors — recheck with the testing tools.

If you are ahead: Your site may have had a stronger starting position than average. Move into weeks 5-6 immediately without waiting.

Weeks 5-6: Content Optimization

Goal: Transform your existing pages into content that AI models actively want to cite.

Time investment: 8-12 hours per week.

Week 5: Rewrite Your Top 5 Pages

Task 17: Add BLUF (Bottom Line Up Front) to each page. 44.2% of AI citations come from the first 30% of a page's content. For each of your five target pages, write a 2-3 sentence summary at the very top that directly answers the main question the page addresses. Do not bury the answer after an introduction — put it first. Our guide on writing content that AI models want to cite covers the BLUF technique in detail.

Task 18: Restructure content into quotable chunks. AI models extract content in fragments of roughly 50-150 words. Go through each page and restructure paragraphs so that individual sections can stand alone as complete answers. Each chunk should: make sense without surrounding context, contain a clear claim or fact, and use specific language rather than vague references like "this" or "as mentioned above."

Task 19: Add definition blocks for key terms. Wherever your content introduces or uses an important concept, add a clear, one-sentence definition formatted as: "[Term] is [definition]." AI models heavily favor content that provides clean definitions — these become the direct answers in AI responses.

Task 20: Update all page metadata. Rewrite title tags and meta descriptions for your top 5 pages with AI retrieval in mind. Include the primary question the page answers in the title. Write meta descriptions as direct answers (not teasers) — AI models sometimes use meta descriptions as the source for short answers.

Week 6: FAQ Sections and Internal Linking

Task 21: Add FAQ sections to your top 5 pages. Write 4-6 relevant questions and answers for each page. Format them as proper FAQ sections with H3 headings for each question. Add FAQPage schema markup to each. The questions should be phrased exactly as users would ask them — not as keyword-stuffed variations.

Task 22: Build internal links between optimized pages. Create contextual links between your top 5 pages and supporting content. Use descriptive anchor text (not "click here") that includes the target page's primary topic. Internal linking helps AI models understand the relationships between your content entities.

Task 23: Add source citations to factual claims. Wherever your content states a statistic, research finding, or factual claim, add a visible source citation. AI models are more likely to cite content that itself demonstrates credibility through sourcing. This is especially important for YMYL (Your Money or Your Life) topics.

Task 24: Test optimized pages in AI platforms. After optimizing each page, wait 3-5 days, then test by asking ChatGPT, Gemini, and Perplexity questions that the page answers. Note whether the AI now cites your content, how accurately it represents your information, and which portions it quotes.

Expected outcome by end of week 6: Your five most important pages are fully optimized with BLUF structure, quotable chunks, FAQ sections with schema, source citations, and internal links. Initial AI citations should start appearing for well-optimized pages.

Weeks 7-8: Off-Site Authority Building

Goal: Strengthen the third-party signals that AI models use to verify your authority and decide whether to cite you.

Time investment: 6-10 hours per week.

This phase is critical because brands are cited 6.5x more often from third-party sources than from their own domains. AI models triangulate your authority by checking how consistently you appear across multiple platforms.

Week 7: Wikipedia, Wikidata, and Knowledge Sources

Task 25: Audit your Wikipedia and Wikidata presence. Search for your brand, key personnel, and primary products on Wikipedia and Wikidata. AI models — especially ChatGPT and Gemini — rely heavily on these as authoritative entity sources. If your brand has a Wikipedia article, verify that all information is current and accurate. If it does not, and your brand meets Wikipedia's notability criteria, begin preparing a draft (note: Wikipedia has strict rules about conflict of interest, so consider working with an experienced editor).

Task 26: Claim and optimize your Wikidata entity. Even if you do not have a Wikipedia article, you can create a Wikidata entry for your organization. Include your official name, founding date, website URL, industry classification, and key personnel. Wikidata is one of the primary knowledge graphs that AI models query for entity information.

Task 27: Update all business directory and knowledge panel listings. Ensure your Google Business Profile, LinkedIn company page, Crunchbase profile, and industry-specific directories all contain identical and accurate information: company name, description, founding date, location, and key personnel. Entity consistency across platforms is a strong trust signal for AI models.

Week 8: Reddit, YouTube, and Review Platforms

Task 28: Build a Reddit presence strategy. Perplexity cites Reddit in a significant portion of its responses, and ChatGPT increasingly references Reddit discussions as social proof. Identify 3-5 subreddits relevant to your industry. Begin contributing genuinely helpful answers (not promotional content) that demonstrate expertise. Over time, your brand will naturally come up in relevant recommendation threads.

Task 29: Optimize YouTube content for AI citation. Perplexity cites YouTube in 16.1% of its responses. If you have YouTube content, optimize video titles as complete questions, write detailed descriptions (500+ words) with key information from the video, add timestamps/chapters, and ensure closed captions are accurate. If you do not have YouTube content, consider creating 2-3 short expert videos answering common industry questions.

Task 30: Manage review platform presence. Claim and respond to reviews on Google, Trustpilot, G2, Capterra, and industry-specific review sites. AI models use review aggregations as authority signals. A business with 50+ reviews and a 4.2+ rating is cited more favorably than one with no review presence.

Task 31: Audit mentions of your brand across the web. Search for your brand name in ChatGPT, Gemini, and Perplexity. If AI models are mentioning you with incorrect information, trace the incorrect data to its source. Often, inaccurate AI answers come from outdated directory listings, old news articles, or incorrect Wikipedia data. Fix the source and the AI will eventually update.

Expected outcome by end of week 8: Consistent entity information across all major platforms, active presence on Reddit and YouTube, accurate knowledge graph entries, and review profiles claimed and active.

Milestone: Day 60 Checkpoint

Run your AI visibility scan again and update your monitoring spreadsheet.

What to expect at day 60:

• AI Score improvement: 20-25 points above your original baseline
• AI citations: Your optimized pages should be appearing in at least 1-2 AI platforms for relevant queries
• GA4 data: Measurable (if small) AI referral traffic, enough to confirm the channel is active
• Content: Top 5 pages fully optimized with BLUF, quotable chunks, FAQ sections, and schema
• Off-site: Wikipedia/Wikidata accurate, directory listings consistent, Reddit and YouTube strategy underway

If you are behind: The most common blocker at day 60 is content that is technically optimized but lacks Information Gain — it says the same things competitors say, so AI has no reason to cite you specifically. Review your top 5 pages and identify what unique data, perspective, or expertise you can add that no competitor offers.

If you are ahead: Some businesses see citations within weeks of their first content optimization. If your AI Score is already 25+ points above baseline, focus weeks 9-10 on scaling what works rather than following the plan rigidly.

Weeks 9-10: New Content Creation

Goal: Expand your AI-visible footprint with new, purpose-built content designed to capture AI citations.

Time investment: 10-12 hours per week.

Week 9: Publish 3-5 AI-Optimized Articles

Task 32: Identify high-opportunity topics. Review the queries you tested in weeks 1 and 6. Which questions did AI models answer poorly or without citing any strong source? These are your content opportunities. Also check: what questions do your customers ask that current AI answers get wrong or answer incompletely?

Task 33: Write and publish 3-5 new articles. Each article should follow the full AI SEO checklist: BLUF opening, quotable chunks, FAQ section with schema, Article schema markup, internal links, source citations, and author bio. Aim for 1,500-2,500 words per article. Focus on depth and accuracy over volume — one well-structured article outperforms five shallow ones.

Task 34: Structure each article for entity recognition. Use clear, consistent terminology. Define key terms explicitly. Reference related entities (companies, tools, concepts) by their proper names. Include comparison sections where relevant — "X vs Y" content is heavily cited by AI because it directly answers comparison queries.

Task 35: Interlink new articles with existing optimized pages. Each new article should link to at least 2-3 of your existing optimized pages, and vice versa. This creates a content cluster that AI models can traverse, increasing the chances that any entry point leads to multiple citations.

Week 10: Content Cluster Strategy

Task 36: Map your content cluster architecture. Identify your core topic (your primary area of expertise) and map 8-12 supporting subtopics around it. Your existing pages and new articles should cover these subtopics. Identify gaps — subtopics that are not yet covered. These become your content pipeline for the next quarter.

Task 37: Create a pillar page (if you do not have one). If your site lacks a comprehensive, authoritative page on your core topic, create one. A pillar page is 3,000-5,000 words, covers the topic broadly, links to every subtopic page, and serves as the "hub" of your content cluster. AI models are more likely to cite content that sits within a clear topical authority structure.

Task 38: Add "Related Questions" sections to all content. At the end of each article, add a section with 3-5 related questions that link to other pages on your site. This mimics the "People Also Ask" pattern and gives AI models additional entry points to your content cluster.

Expected outcome by end of week 10: 3-5 new AI-optimized articles published, a content cluster strategy documented, and a content pipeline for the next quarter identified.

Weeks 11-12: Monitor and Iterate

Goal: Measure the full impact of your 90-day effort, identify what worked best, and plan the next phase.

Time investment: 3-5 hours per week.

Week 11: Comprehensive Analysis

Task 39: Run a full AI visibility re-scan. Get your updated AI Score from AImetrico and compare it to your day-1 baseline and day-30/day-60 checkpoints. Document the trajectory.

Task 40: Test all target queries across AI platforms. Repeat the same queries from week 1 across ChatGPT, Gemini, Perplexity, and Claude. For each query, note: are you mentioned (yes/no), is the information accurate, are you cited with a link, and which competitors appear alongside you.

Task 41: Analyze GA4 AI referral data. Pull 90 days of AI referral traffic data. Identify: total AI referral sessions, month-over-month growth, which AI platforms send the most traffic, which pages receive the most AI referral visits, and the conversion rate of AI referral traffic compared to organic search.

Task 42: Run competitor analysis. Pick your top 3-5 competitors and check their AI visibility: ask AI platforms the same questions and note who gets cited. Check their robots.txt, schema markup, and llms.txt. Identify where competitors are ahead and where you have an advantage. For a structured approach, refer to our guide on building an AI SEO strategy from scratch.

Week 12: Iterate and Plan

Task 43: Identify your highest-performing optimizations. Which changes had the biggest impact? In most cases, the ranking will be: robots.txt fixes (immediate impact), content restructuring with BLUF (citations within days), FAQ schema (improved AI interpretation), and off-site authority (gradual but compounding).

Task 44: Fix underperforming pages. If any of your top 5 optimized pages are still not being cited, diagnose the issue. Common causes: content lacks unique information (Information Gain), page speed regressed, schema has validation errors, or the topic is dominated by extremely authoritative competitors (Wikipedia, government sites).

Task 45: Plan your next quarter. Based on 90 days of data, create a next-quarter plan that includes: a content calendar (2-4 new articles per month), a list of pages to refresh and re-optimize, off-site authority actions to continue, and a target AI Score for day 180.

Task 46: Set up automated monitoring. Move from manual weekly checks to a more sustainable system. Set up alerts for brand mentions in AI platforms, schedule bi-weekly AI Score scans, and create a monthly AI SEO report template that tracks Score, citations, referral traffic, and competitor position.

Expected outcome by end of week 12: A comprehensive "before and after" report, a clear understanding of what works for your site, and a documented plan for the next quarter.

Milestone: Day 90 Final Review

This is the moment of truth. Here is what a successful 90-day execution looks like:

| Metric | Starting Point (Day 0) | Target (Day 90) | |---|---|---| | AI Score | 10-20 (typical starting range) | 50-65 (25-35 point increase) | | AI platform citations | 0-1 platforms mention you | 2-4 platforms cite you regularly | | Schema coverage | Little or no structured data | Organization, Article, FAQ, Person on all key pages | | robots.txt | AI crawlers blocked or misconfigured | All search bots allowed, training bots selectively managed | | Content structure | Traditional SEO-style content | BLUF, quotable chunks, FAQ sections on top 10+ pages | | Off-site presence | Inconsistent or absent | Consistent entity data across 10+ platforms | | GA4 AI referrals | Not tracked | Tracking active, baseline data established | | llms.txt | Does not exist | Published and validated | | Content cluster | No AI-optimized content | 8-10+ AI-optimized pages in a structured cluster |

Remember: AI SEO is not a 90-day project — it is an ongoing discipline. The 90-day roadmap establishes your foundation. The compound effect of consistent execution means that months 4-6 will typically show faster improvement than months 1-3, because AI models have had more time to crawl, index, and build confidence in your content.