GEO checklist: 40 checks to get your brand named in AI answers (2026)

How to use this checklist

This is a working GEO checklist, not a theory piece. Generative engine optimization (GEO), also called answer engine optimization (AEO) or AI SEO, is the work of getting your brand named, early and accurately, when buyers ask AI assistants for recommendations. The checks below are grouped in the order we would actually do them: measure first, fix positioning, then work on the pages and sources that answers lean on, then re-measure.

Score each check as done, partly done or not started. Do not try to finish all 40 in a week. Most teams get the biggest lift from the first two sections and a handful of comparison pages, and only then need the rest. If you want the reasoning behind the checks, read our guides on what generative engine optimization is and on answer engine optimization; this page assumes you already know why the work matters and want to know what to do.

Everything here can be done with a spreadsheet and a few hours a week. None of it requires paid placement, because organic AI answers do not sell placement.

What our own data shows

We ran unbranded buyer questions ("best [category]", "[competitor] alternatives", "X vs Y") through AI search for 50 software brands. These are the only numbers we quote anywhere on this site, and we quote them as they are.

• 39 of 50 brands (78%) were never named in any answer to their own category's buyer questions.
• All 30 brands that had launched in the previous month scored 0.
• Being known is not enough: an established e-signature tool scored 52/100 and an AI voice tool scored 37/100.
• In 33 of 50 cases the most cited source was a competitor's own website: its "vs" pages, docs and help centre.
• Review sites were the most cited source for only 6 of 50 brands.

Section 1: Baseline and measurement (checks 1 to 6)

1. 01Write 10 to 20 unbranded buyer prompts for your category: "best [category] for [audience]", "[market leader] alternatives", "[A] vs [B]" and two or three with constraints such as price, team size or an integration.
2. 02Remove your own brand name from every prompt. Branded prompts measure whether people who already know you can find you, not whether new buyers discover you.
3. 03Run each prompt more than once and record every brand named, in order. Answers vary between runs, so one screenshot is not a baseline.
4. 04Record every source cited or linked, then count how often each domain appears across the whole set.
5. 05Calculate three numbers: mention rate (share of answers that name you), average position when named, and share of voice (your mentions divided by all brand mentions).
6. 06Save the prompt set with a date and do not change it for at least a month, so later results are comparable.

Section 2: Positioning and category clarity (checks 7 to 12)

1. 01Write one sentence of the form "[Brand] is a [category] for [audience]". Use an existing category name buyers already search for, not a coined one.
2. 02Put that sentence, or a close variant, in your title tag, H1 and meta description.
3. 03Use the same category words on your about page, pricing page and documentation home.
4. 04Update every directory listing, marketplace profile and social bio to match the same sentence.
5. 05Check that your homepage's first visible paragraph says what you are without needing images or scripts to load.
6. 06Remove or soften taglines that could describe any product ("the platform for modern teams"). Assistants match brands to categories, and vague copy makes that match harder.

Section 3: Pages that answer buying questions (checks 13 to 20)

1. 01Publish an honest "[You] vs [main competitor]" page with a comparison table built from facts: prices, limits, key features.
2. 02Add a short "Choose them if, choose us if" section to every comparison page. Saying where a competitor is stronger makes the rest of the page credible.
3. 03Publish a "[Market leader] alternatives" page that lists five to eight real options, yours included, with who each suits.
4. 04Make pricing public, with numbers, on a crawlable page. When answers compare cost, products without a stated price are easy to leave out.
5. 05Make core documentation and help articles readable without a login. Gated pages cannot be read or cited.
6. 06Add FAQ sections that answer real buying questions (pricing, limits, integrations, who it is for) on your key pages.
7. 07Date your comparison and pricing pages and review them every quarter. Stale prices undermine trust in the whole page.
8. 08Cover at least one constraint prompt where a smaller product can win, such as "cheapest [category] for a 5-person team" or "[category] with [specific integration]".

Section 4: Technical readability (checks 21 to 27)

1. 01Add Organization structured data with your name, logo, URL and sameAs links to your official profiles.
2. 02Add SoftwareApplication or Product structured data with an accurate category and offers that include prices.
3. 03Mark up genuine FAQ sections with FAQPage structured data. Do not mark up questions nobody asked.
4. 04Publish an llms.txt file at your site root: a one-line description, then links to pricing, docs, comparisons and changelog with a short note on each.
5. 05Server-render key copy so crawlers that fetch pages on demand see your category, pricing and features without running scripts.
6. 06Keep a clean XML sitemap with canonical URLs and submit it to the major search consoles, since many assistants use web search results behind the scenes.
7. 07Check robots.txt. If you want to be cited, do not block reputable search and AI crawlers from the pages you want cited.

Section 5: Sources you do not own (checks 28 to 35)

This is the section most teams skip, and in our data it is often the one that decides the answer. In 33 of the 50 brands we checked, the most cited source was a competitor's own website. Review sites were the top source for only 6. Work from your source-frequency list from check 4, not from assumptions.

1. 01Take the three most cited third-party pages for your category and note whether you are mentioned on each.
2. 02Email the author of each cited roundup with a short, honest note on who your product is best for, plus a demo account if relevant. Ask for accuracy, not a ranking.
3. 03Ask active, happy customers for recent reviews on the review platforms that actually appear in your answers, not on every platform.
4. 04Answer real questions in the community threads that get cited, and disclose that you work on the product.
5. 05Complete listings on the directories and launch sites relevant to your category, using the category sentence from check 7.
6. 06Correct factual errors about you on third-party pages: outdated prices, discontinued plans, wrong category.
7. 07Look at competitors' comparison pages that mention you. If they misstate your pricing or features, publish your own accurate comparison so answers have a second source.
8. 08Track which cited sources you have won. This is a leading indicator: it usually moves before mention rate does.

Section 6: Re-measure and iterate (checks 36 to 40)

1. 01Re-run the exact prompt set from check 6 on a fixed cadence, weekly or monthly.
2. 02Compare mention rate, average position and share of voice against the baseline, and treat changes of a few points on one run as noise.
3. 03Check description accuracy: when you are named, are you described as what you are and for whom?
4. 04Update the source list. If a new domain has become the most cited source, add it to your outreach plan.
5. 05Pick the next two or three checks with the biggest gap, ship them, and repeat. Expect movement over weeks as cited pages change and are re-crawled, not overnight.

What to prioritise if you only have one week

If your mention rate is near zero, which is common, do not start with technical polish. In our set, all 30 brands launched in the previous month scored 0, and 39 of 50 brands were never named at all. For a brand at zero, the order that makes sense is: checks 1 to 6 to see which sources matter, checks 7 to 10 so every source can describe you consistently, checks 13 to 16 so there is at least one page that answers the comparison buyers ask about, and checks 28 and 29 for the single most cited roundup.

If you are already named sometimes, focus on position and accuracy instead. Read how you are described, fix the category sentence where it drifts, and target the constraint prompts (check 20) where you appear once and could appear consistently.

Common checklist mistakes

• Ticking boxes without a baseline. Without checks 1 to 6 you cannot tell which work moved anything.
• Changing the prompt set every week. A moving target makes every trend meaningless.
• Writing generic blog posts instead of comparison and alternatives pages. Answers to buying questions draw heavily on pages that compare options directly.
• Self-declared rankings such as "the #1 tool for X". Independent sources carry more weight than claims about yourself.
• Assuming fame is enough. In our data an established e-signature tool scored 52 out of 100 and an AI voice tool 37.
• Treating llms.txt or schema as a ranking switch. They help assistants describe you accurately; third-party mentions do most of the work of getting you named.

Using this checklist with clients

Agencies can turn the 40 checks into a client scorecard: run the baseline, score each check, and propose a fixed-fee sprint covering the largest gaps. Re-running the same prompts each month gives clients a before-and-after view tied to specific changes, which is easier to sell than a promise of rankings. Our GEO for agencies page covers packaging and pricing in more detail.

Where ShownInAI fits

You can do every check here by hand. If you want the measurement part done for you, the $49 report runs 10 unbranded buyer prompts for your domain and returns a 0 to 100 score, the competitors named instead of you, the cited sources and the first fix. The $299 GEO audit covers 20 queries, a competitor comparison and a 30-day fix plan. Tracking plans start at $29 a month and re-run the same prompt set on a schedule, which covers checks 36 to 40.