The 90-Day AEO Playbook: A Week-by-Week Plan for Your First Quarter

Ninety days is enough time to go from invisible to cited, if you spend them in the right order. Most brands that try AEO on their own skip straight to writing content, skip the technical audit that would have told them AI crawlers were blocked the whole time, and wonder three months later why nothing changed. This playbook fixes the sequencing problem. It gives you a week-by-week plan broken into three phases: audit and technical fixes, content and entity building, and measurement and expansion.

I built this plan the same way I built the AEO Maturity Model: from working the same problem across client after client and noticing the same mistakes. Brands skip the audit. They write content before fixing crawler access. They never define what “success” looks like, so ninety days pass and nobody can say whether the work moved the needle. This playbook removes the guesswork by telling you exactly what to do each week and why that week comes before the next one.

Why sequence matters more than effort

You can pour a hundred hours into content and still get zero AI citations if your robots.txt blocks GPTBot. You can build a flawless entity graph and still lose to a competitor whose content simply answers the question faster. AEO has dependencies. Technical access has to come before content matters. Content has to exist before entity signals can reinforce it. Measurement has to start early enough that you have a baseline to compare against by week twelve.

This is the core idea behind the phased structure. Each phase builds the floor the next phase stands on. Skip a phase or do them out of order, and you spend your ninety days fixing problems you should have caught in week one.

Before you start: set your baseline

Pick 10 to 15 queries that represent how real prospects would ask an AI model about your category. Not your brand name, the actual questions a buyer types. “Best CRM for a 10-person sales team.” “How much does a kitchen remodel cost in Phoenix.” “Who does AEO for local service businesses.” Run every query through ChatGPT, Perplexity, and Google AI Overviews, and record whether your brand appears, which competitors appear instead, and what kind of source gets cited.

This baseline is what makes week twelve meaningful. Without it, you have no way to prove the work did anything. Save the results in a simple spreadsheet: query, engine, cited or not, competitor cited instead, date checked. You will run this same check every week of the ninety days.

Being “cited” means something specific here. It is not the same as appearing somewhere in a long response. Look for three things: your brand named directly in the answer, a link back to your site included as a source, or your content quoted or closely paraphrased. A passing mention buried in a list of six options counts for less than a direct recommendation. Track both so you can see whether you are moving from passing mentions toward direct recommendations over the ninety days, rather than only watching whether the count of appearances goes up.

Who should run this plan

This plan assumes one person owning it end to end, even if the actual writing, development, and outreach get split across a team. Someone with access to your CMS and a developer who can touch robots.txt and schema markup covers most of phase one. Phase two needs a writer who understands the content and someone with entity-building patience, since directory cleanup and Wikidata entries are tedious rather than difficult. Phase three needs whoever owns the reporting relationship, because week twelve is where the work gets summarized and defended.

Budget four to six hours a week if you are doing this alongside a full-time role. Weeks 2, 3, and 7 run heavier because of the technical and entity groundwork. Weeks 9 and 12 are lighter on execution and heavier on analysis. If you are running this for a client rather than your own brand, expect the entity-building weeks to take longer, since directory access and profile ownership often sit with someone outside the marketing team.

Phase 1: Audit and technical fixes (weeks 1-4)

The first phase is about clearing obstacles, not creating anything new. Every day spent here removes a barrier that would otherwise cap the return on every hour you spend in phases two and three.

Week 1: Full technical and content audit

Start by checking your robots.txt file for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended. Confirm none of them are blocked, intentionally or through an inherited template nobody reviewed. Next, run your homepage and your five most important pages through a schema validator and note every error. Then check whether your site renders content in the initial HTML or requires JavaScript, since AI crawlers frequently see an empty shell on client-side-only sites.

While the technical checks run, inventory your existing content. List your top 20 pages by traffic or business importance and score each one against a simple checklist: does it lead with a direct answer, does it have a named author, does it include an FAQ section, is it up to date. This inventory becomes your worklist for phase two, so do not skip it even if it feels premature this early.

Close week one by checking your entity presence too, even though the heavier entity work does not start until week seven. Search your brand name and note whether a Knowledge Panel appears, whether a Wikidata entry exists, and whether your name, address, and phone number match across your top ten directory listings. You are not fixing anything yet. You are building the full picture so nothing surprises you later in the plan.

Week 2: Fix crawler access and core technical gaps

Unblock any AI crawlers you found restricted in week one. This is usually a five-minute robots.txt edit, but it does more for your AI visibility than almost anything else in the entire ninety days. If your site is JavaScript-rendered and crawlers cannot see your content, prioritize server-side rendering or pre-rendering for at least your highest-value pages, even if a full fix takes longer than this week.

Add or repair your XML sitemap and confirm it is referenced correctly in robots.txt. Run a page speed check on your top pages and flag anything scoring below 80 on mobile, since AI crawlers operate on time budgets and slow pages get skipped.

Week 3: Schema markup rollout

Implement or repair Organization schema on your homepage, including your logo, founding information, and any social profiles you want connected. Add Article schema to your existing blog content and FAQPage schema to any page that already has question-and-answer content, even informally. If you have named subject matter experts, add Person schema for each of them and link it back to the Organization schema.

Validate everything you implement. A broken schema block is often worse than no schema block, because it signals inconsistency rather than absence.

Use Google's Rich Results Test on every page you touch this week, including pages beyond the homepage. It is common to fix schema on a handful of flagship pages and leave dozens of older posts with outdated or duplicate markup from a previous CMS migration. Duplicate Organization schema across multiple pages, each with slightly different details, confuses more than it clarifies. Pick one canonical version, reference it with an @id, and point every other page back to that single source.

Week 4: Create your llms.txt and close out phase one

Add an llms.txt file to your site root with a clear, concise description of your organization and links to your most important pages. This is a newer standard, and adoption alone puts you ahead of most competitors who have not heard of it yet. Spend the rest of the week re-running your baseline queries and technical checks to confirm every phase-one fix actually took effect. Document what changed so you have a clean record before content work begins.

Phase 2: Content and entity building (weeks 5-8)

With the technical floor in place, phase two shifts to the two pillars that determine whether AI models actually want to cite you: content structure and entity recognition. This is where most of the visible work happens, and where the AEO content checklist becomes the day-to-day reference for anyone writing or editing.

Week 5: Reformat your highest-value existing pages

Before writing anything new, go back to the content inventory from week one and reformat your top 10 pages. Rewrite the opening paragraph of each to lead with a direct answer instead of a warm-up introduction. Add a definition box for any concept the page introduces. Convert any comparison currently buried in prose into an HTML table. This reformatting work is fast, and because these pages already have authority and traffic, the AI visibility gain shows up quickly.

Week 6: Build FAQ sections and named authorship

Add FAQ sections to every page that does not have one, five to eight questions each, phrased the way a person would actually ask an AI model. Match the FAQPage schema you added in week three so the visible content and the structured data stay in sync. If your content is attributed to generic bylines like “Staff” or “Admin,” replace them with real names and build out author bio pages with actual credentials. Named, verifiable authorship is a signal AI models weigh, and it costs nothing but the willingness to put a name on the work.

Write the FAQ answers the same way you would answer a colleague who asked you the question directly. Two to four sentences, no hedging, no restating the question back at the reader. “How much does a kitchen remodel cost” gets a real number or range in the first sentence, not a paragraph about how costs vary. AI models extract the most direct answer available, so a vague FAQ answer loses to a competitor's specific one even when your underlying expertise is stronger.

Week 7: Start entity building

Claim or create your Google Business Profile if you have not already, and check your name, address, and phone number consistency across every directory and listing you can find. Create a Wikidata entry for your organization if one does not exist. Set up or clean up your LinkedIn Company Page and Crunchbase profile, and make sure your Organization schema includes sameAs links back to every profile you control.

Run the ChatGPT test from the maturity model: ask ChatGPT directly, “What is [your brand]?” and “What does [your brand] do?” If the answer is wrong, outdated, or nonexistent, that gap is exactly what this week's entity work is meant to close over time.

Week 8: Publish new cornerstone content

With reformatting and entity groundwork underway, write one new piece of cornerstone content targeting a query you identified as uncited in your baseline. Apply the full structure from day one: direct answer opening, definition box, FAQ section with matching schema, and internal links to your service pages and other articles. One well-built page targeting a real gap outperforms five thin posts written to fill a content calendar.

Phase 3: Measurement and expansion (weeks 9-12)

The final phase shifts from building to verifying and compounding. You have technical fixes and initial content and entity work behind you. Now you find out what worked, fix what did not, and start expanding into new territory.

Week 9: Re-run the full baseline

Query your same 10 to 15 target queries across ChatGPT, Perplexity, and Google AI Overviews again, using the identical method from your original baseline. Compare directly: which queries now cite you that did not before, which are unchanged, which still favor a competitor. This is the first hard data point on whether the first eight weeks worked, and it should shape everything in the remaining three weeks.

Week 10: Diagnose the gaps

For queries still not citing you, dig into why. Pull up the content that is winning instead and compare it against yours directly. Is it more comprehensive? Does it have stronger entity signals? Is it simply better formatted for extraction? Sort your gaps into the four pillars from the maturity model so you know whether the fix is another content pass, a schema fix, or more entity work. Brands that skip this diagnostic step tend to just write more content, even when the actual blocker was technical or entity-related.

Pay particular attention to queries where a competitor appears that was not in your baseline results at all. A new entrant showing up in week nine means someone else is actively working their AEO, and it is worth a closer look at what changed on their site in the same window you were working on yours. Sometimes the answer is a single piece of content they published. Sometimes it is a press mention that boosted their entity signal. Either way, this comparison tells you what the competitive bar actually looks like right now, not what it looked like when you set your baseline.

Separate the queries where you are close from the queries where you are far behind. A query where you are mentioned but not linked needs a different fix than a query where your brand never appears at all. Triage this way so week eleven's limited hours go toward the fixes most likely to move a query from “mentioned” to “cited as the primary source,” rather than spreading effort evenly across every gap regardless of how close it already is.

Week 11: Targeted fixes and a second cornerstone piece

Apply the fixes identified in week 10. If the gap was formatting, go back and add the missing tables, lists, or FAQ sections. If the gap was entity related, push harder on directory consistency and third-party mentions. If a genuine content gap remains, write a second cornerstone piece targeting the specific angle where you are losing. This week is about closing the exact gaps your data pointed to, not adding volume for its own sake.

If you have room left in the week after the priority fixes, use it on outreach rather than more internal edits. A single guest post, podcast appearance, or industry directory listing secured this week can move an entity gap faster than another round of on-site changes. This is also a good week to pitch a journalist or industry publication if you have original data or a genuine point of view worth covering, since third-party mentions compound with everything else you have built over the previous ten weeks.

Week 12: Report, plan, and set up ongoing monitoring

Close out the quarter with a clear before-and-after comparison: baseline citations versus week-twelve citations, technical issues fixed, schema types implemented, entity signals added, and content pages reformatted or created. Set up a recurring weekly or biweekly check of your target queries so monitoring continues past the ninety days instead of stopping the moment the formal plan ends. Use what you learned to draft your priorities for the next quarter, since AEO compounds and the second ninety days should build directly on the first.

If you are reporting this work to a client, a boss, or a partner, resist the urge to present only the wins. A query that stayed flat despite real effort is useful information, not a failure to hide. Note which pillar is holding that query back and what you plan to do about it in quarter two. A report that shows both the movement and the remaining gaps reads as more credible than one that only shows good news, and it sets honest expectations for what the next ninety days need to cover.

Common pitfalls during the 90 days

A few mistakes show up often enough across clients that they are worth naming directly, so you can watch for them instead of discovering them in week ten.

Treating week one as a formality

Teams eager to see progress sometimes rush the audit to get to the “real” work. This backfires. A blocked crawler or a broken schema block discovered in week eight means redoing content work that should have been built on a fixed foundation from the start. The audit is not overhead. It is the step that determines whether the rest of the plan can succeed at all.

Checking citations inconsistently

Running your baseline queries once in week one and once in week nine, with nothing in between, tells you less than you think. AI models update frequently, and a query that cites you in week four might not by week six. Weekly checks, even quick ones, catch swings early and give you a real trend line instead of two disconnected snapshots.

Writing content before fixing the byline

New cornerstone content published under a generic “Team” byline undercuts itself before it has a chance to earn citations. Fix authorship in week six, before week eight's new content goes live, so the new piece launches with a credentialed author already in place rather than retrofitting it later.

Confusing entity work with a one-time task

Claiming a Google Business Profile or creating a Wikidata entry feels like a checkbox you complete once. In practice, directory listings drift, NAP details get overwritten by third-party aggregators, and profiles go stale. Entity work started in week seven needs a light recheck built into your week twelve report, and ideally a quarterly recheck after that.

Tools you actually need

You do not need a specialized platform to run this plan. A schema validator, a robots.txt checker, a spreadsheet, and access to ChatGPT, Perplexity, and Google directly cover most of the ninety days. Google's Rich Results Test and the Schema Markup Validator handle phase one's structured data checks. Screaming Frog or a similar crawler catches broken sitemaps and JavaScript rendering issues at scale if your site has more than a handful of pages.

For citation tracking, the manual approach described throughout this plan works fine at the query volumes involved here, ten to fifteen queries checked weekly across three engines. If you manage AEO for multiple brands or need to track citations at a larger scale, dedicated AI citation tracking tools exist and can automate the weekly query runs, but they are not required to complete this plan. The habit of checking consistently matters more than the tool you use to do it, and a disciplined spreadsheet run every week for ninety days will beat an expensive dashboard checked once a month.

What to skip if you have a smaller team

If you do not have the bandwidth to run all three phases fully staffed, do not compress the timeline. Extend it instead, and protect phase one no matter what. A slower rollout that keeps the technical foundation intact will outperform a rushed ninety days that skips crawler checks or schema validation to hit an arbitrary deadline. The order of operations matters more than the speed.

If you have to cut scope somewhere, cut the volume of new content in phase two before you cut the reformatting work. Reformatting existing pages is faster, cheaper, and often produces a bigger visibility gain than a brand-new article, because the page already carries traffic and authority.

How this plan connects to the broader AEO picture

This ninety-day plan is the execution layer. For the strategic context behind why each phase matters, our Complete Guide to AEO covers the full picture of how AI answer engines choose sources and why entity signals and structured content matter as much as they do. And once you have run this plan once, the AEO Maturity Model gives you a way to score where you landed across all four pillars and decide what quarter two should prioritize.

Ninety days will not make you Level 5. It will make you visible, measurable, and positioned to compound from there. That is the entire point of a first quarter: not to finish the work, but to build a foundation the rest of the work can stand on.