AEO tools are software and free reports that help a business get picked as the answer when someone asks a question in a search engine or an AI assistant. Five kinds exist, defined by the job: question research, answer formatting and snippet eligibility checks, structured data validation, answer tracking across assistants, and accuracy monitoring of what assistants say about your brand.
The pages that rank for a search on AEO tools mostly list trackers. A tracker is one of the five jobs, and it is compared tracker by tracker in AI visibility tools compared: 12 tools. This guide covers the whole workflow, and ends with what Google's own documentation says today about FAQ markup, which changed in 2026.
I opened the vendor or documentation page for each of 18 tools and instruments on 21 November 2026 and recorded what each says about method, limits and cost. When a page was silent, the text says "not published". Prices appear only where I read them on the vendor's own page that day. I do not rank any tool and I have not run any as a customer. I sell a related service, described in GEO agency, consultant or tool, so read the groupings as a map, not a recommendation. For the wider set around crawlers, content and press, read the companion guide to GEO tools.
#The short version
| Question | Answer from the pages read on 21 Nov 2026 |
|---|---|
| How many jobs do AEO tools do? | 5, in workflow order: research, formatting, markup, tracking, accuracy |
| Can you mark up a page to win a featured snippet? | No. Google says "You can't" and that its systems decide |
| Does FAQ markup still produce a rich result in Google? | No. Google says it stopped appearing from 7 May 2026 and removed the documentation on 15 June 2026 |
| Does Google require special markup for AI Overviews or AI Mode? | No. Its AI features page says no special schema.org structured data is needed |
| Which accuracy tools publish an accuracy rate for their own checks? | None of the 2 dedicated ones, Profound FactCheck and seoClarity ArcAI Accuracy |
| Which free tools validate markup? | Rich Results Test and Schema Markup Validator |
#Group 1: Question research (what people ask)
The job. Find the exact questions your buyers type or speak, so each page answers one of them directly. This comes first because every later job depends on the question list.
Tools, from their own pages:
- AlsoAsked. Provides live, geo-specific People Also Ask questions. The page does not describe its data source beyond reference to Google's People Also Ask feature. Plans read on 21 November 2026, in USD plus tax: Basic $9 a month for 100 credits and 24 hours of history, Lite $17.25 for 300 credits and 1 month, Pro $35.25 for 1,000 credits and 1 year. All plans list PNG and CSV export, bulk searches, API access and all regions and languages. No free tier is mentioned (AlsoAsked pricing).
- AnswerThePublic. Describes itself as a way to uncover the questions people ask across platforms, with a newer AI Dashboard showing how AI describes a keyword. The page does not name a data source. It mentions a free daily limit under a fair usage rule and does not state the number. Plans billed yearly: Starter US$13.33 a month for 100 searches, Growth US$66 for 200, Business US$132.66 for 300. A banner on the page names a "top plan" of US$1,592 that does not match the table and is not explained (AnswerThePublic pricing).
- Google Search Console. Your own query data, not a sample of someone else's. The Performance report supports a custom regex filter on queries and pages using RE2 syntax, with a "doesn't match" option, so you can isolate queries that start with who, what, how, why or contain "vs". Google notes that filtering by query or URL can make totals differ from other views because of data truncation and omitted anonymized queries (Search Console Help). Google's help page on the report states the default view covers the past three months (Search Console Help, Performance report).
- Google Trends. Shows interest in search terms and topics by time, location and popularity. It shows relative interest, and the page does not list question data (Google Trends).
- Profound Prompt Volumes. Described on the pricing page as showing what users prompt AI answer engines, with history for all time, CSV or JSON export and API access. It is part of the Enterprise plan, which has custom pricing, and how volumes are estimated is not published (Profound pricing).
Limits. People Also Ask and search-box tools show Google-style queries. AI prompts tend to be longer and more conversational, and only one tool here, Profound, says it measures them, without stating its method. Bing's AI Performance report publishes grounding queries, the phrases an AI used when retrieving your content, which Microsoft says are a sample and which Bing is still refining (Bing Webmaster Blog).
Free route. Search Console's regex filter on your own queries plus Bing's grounding queries gives you real questions at no cost. Add the questions your sales calls and support inbox produce. Those are the ones no tool can sell you.
#Group 2: Answer formatting and snippet eligibility checks
The job. Make sure a page can be quoted: a direct answer near the top, a heading that matches the question, text that a crawler can read, and no controls that block snippets.
What Google says, in its own documentation. The featured snippet page states that you cannot mark up a page to become a featured snippet. Google's systems decide whether a page would make a good one and elevate it. The controls that exist are opt-outs: nosnippet stops featured and regular snippets and takes precedence over data-nosnippet, which excludes marked text, and a lower max-snippet value makes featured snippets less likely but does not guarantee exclusion. A page needs enough text to produce a useful snippet, and Google publishes no minimum length (Google, featured snippets). For AI features, Google says a page must be indexed and eligible to appear with a snippet, that there are no additional requirements for AI Overviews or AI Mode, and that important content should be in text (Google, AI features).
Tools, from their own pages:
- URL Inspection tool. Google's AI features page says to use it to confirm that snippet controls are visible to Googlebot, then allow days to months for a recrawl or request one (Google, AI features).
- Surfer. A content editor with real-time guidelines it describes as competition-based, plus an AI tracker. Plans listed as billed yearly: Discovery $49 a month, Standard $99, Pro $182, Peace of Mind $299, and AI Search Analytics $82. The scoring method is not published beyond "over 500 web and AI signals". A free AI visibility audit is listed (Surfer pricing).
- Frase. Starter $39 a month billed yearly ($49 monthly) with 10 articles and 50 audit pages a month. Professional $103 ($129) and Scale $239 ($299) add AI crawler monitoring, Perplexity tracking, Claude and Gemini tracking and more audit pages. A 7-day trial with no card is listed (Frase pricing).
- Screaming Frog SEO Spider. Its page lists custom extraction using XPath, CSS path or regex, which you can use to check at scale whether every article has an answer paragraph, a question heading or a date. The free version crawls 500 URLs, and the licence was listed at 199 pounds a year. The page contradicts itself on whether custom extraction is available in the free version, so check the download (Screaming Frog).
Limits. No formatting tool can promise a snippet or an AI citation, and Google says so. Content scores are tool-specific and none of the pages published a validated link between a score and being quoted.
Free route. Read each page with the question in mind: does the first two sentences answer it? Is the answer text, not an image? Does the heading match how the question is asked? Then check the page with URL Inspection and look for stray nosnippet or max-snippet rules.
#Group 3: Structured data validation
The job. Confirm that markup parses, matches the visible page, and meets the requirements of the Google feature you target.
Rules from Google. Markup should only describe content users can see. JSON-LD is recommended, and Microdata and RDFa are supported. Pages must stay crawlable. Required properties must be present for rich result eligibility. Valid markup does not ensure a rich result will appear, and a violation can remove rich result eligibility without affecting ranking (Google, structured data policies). Google says all 3 formats are equally fine when valid (Google, structured data intro).
Tools, from their own pages:
- Rich Results Test. A Search Console tool that tests a public URL or pasted code and shows which rich results the markup can generate. The page does not list supported types and points to the structured data gallery (Rich Results Test). Google's intro page says it can also preview how a feature looks in some cases (Google, structured data intro).
- Schema Markup Validator. At validator.schema.org. The announcement says it checks that JSON-LD, RDFa and Microdata are used in widely understood ways and does not check against any specific product's requirements (Schema.org announcement). The Schema.org documentation says it takes a URL or pasted markup, picks up JavaScript-injected data, and does not fetch or interpret other JSON-LD
@contextURLs (Schema.org documentation). The tool's own page does not describe its scope. - Search Console rich result reports. Google tells you to check the rich result status reports after deployment, because templating or serving problems can break markup that previously validated. URL Inspection confirms Google found the markup (Google, structured data intro).
- Screaming Frog SEO Spider. Its page says it can "Extract & validate structured data against Schema.org specifications and Google rich result features", free up to 500 URLs, 199 pounds a year for a licence on the day I read it. The page does not confirm JSON-LD by name (Screaming Frog).
Limits. A validator confirms syntax and, for Google's tool, feature requirements. It does not tell you whether an assistant uses the markup. Google says AI features need no special schema (Google, AI features), and I did not find a vendor page that published a test showing schema changes an AI answer.
Free route. The first two tools above show no price on their pages. Run the two validators on one template per page type, not one page at a time, then use Search Console to watch for regressions.
#What changed for FAQ markup, from Google's own documentation
This section uses only Google Search Central pages opened on 21 November 2026. The URL for the FAQ structured data documentation returned Google's documentation updates log, which carries the dated entries below, and I confirmed them on the updates page itself (Google Search Central, FAQ page URL; Google Search Central, documentation updates).
| Date | Entry in Google's log |
|---|---|
| 14 Sep 2023 | FAQ documentation updated to say the feature is only shown for well-known, authoritative government and health websites |
| 22 Jul 2021 | Hidden-content guidance: the user must be able to access the answer on the page, and expandable sections are acceptable |
| 8 May 2026 | Deprecation notice added: the feature will no longer appear in Google Search starting 7 May 2026 |
| 15 Jun 2026 | FAQ rich result documentation removed |
| 24 Mar 2026 | More supported properties added for Discussion Forum and QA Page markup |
Three consequences follow, each limited to what the documents say.
- The FAQ rich result is gone from Google Search. The log states it will no longer appear from 7 May 2026. The log entry is dated 8 May, a day after the effective date, and the log does not explain the gap.
- The gallery does not list it. The structured data feature gallery I opened lists 25 features, including Q&A, Education Q&A and Discussion forum, and does not list FAQ or mention any feature being removed (Google, structured data gallery).
- QAPage is not a substitute. Google's QAPage documentation says the markup is for a page with one question followed by its answers, where users can submit answers. It lists an FAQ page written by the site itself with no way for users to submit alternative answers as an invalid use (Google, QAPage).
What the documents do not say: whether existing FAQ markup should be removed, whether Bing or any AI assistant reads it, and whether it affects AI Overviews. Google's AI features page says no special schema.org markup is needed for those features. The visible question and answer text on your page is covered by Google's content rules, which require that users can access the answer on the page. If you keep FAQ markup, keep it matching the visible text, per the structured data policies above.
#Group 4: Answer tracking across assistants
The job. Run your question list through assistants on a schedule and record whether you are named or cited. Tracker-by-tracker detail, with allowances and gaps, is in the 12-tool comparison, so this section stays short.
Tools read for this guide:
- Otterly. Lite 29 euros a month for 16 prompts and 4 of 7 engines, Standard 189 euros for 120 prompts, Premium 489 euros for 403 prompts. A credit is one prompt on one engine once, and Claude costs 3 credits per check (Otterly pricing).
- Peec AI. Starter $95 a month for 50 prompts, 3 of 8 models and 4,500 answers a month. Currency not named on the page (Peec AI pricing).
- Ahrefs Brand Radar. Custom Prompts from $50 a month, with a daily, weekly or monthly cadence per platform, and an Index from $199 a month. Mentions and citations reported separately (Ahrefs Brand Radar).
- Semrush AI Visibility Toolkit. Runs queries through ChatGPT and Google's AI Mode. Prompt Tracking is daily, Brand Performance weekly (Semrush knowledge base).
- AthenaHQ. A free Essentials plan with 300 credits across 5 platforms, and Starter at $295 a month for 3,600 credits across 11 models. One credit is one AI response (AthenaHQ pricing).
Limits. Method is the weak point across all five pages: repeats per prompt are not published, and only Peec's and Otterly's pages show an answer allowance or credit rule. Size the plan before you buy: prompts times assistants times days. Fifty prompts on 3 assistants daily is 4,500 answers a month.
Free route. Manual testing: 10 to 20 questions, 5 runs each, fresh session, a sheet recording named brands and cited URLs. My own free check, at /ai-visibility-check, is a done-for-you service of mine. It is not one of the tools in this guide and I have not ranked it against them.
#Group 5: Accuracy monitoring
The job. Being named is not the same as being described correctly. Accuracy tools extract the claims an assistant makes about you and compare them with a source of truth. Semrush's own guidance says frequent mentions do not indicate accuracy, so the two need tracking separately (Semrush blog).
Tools, from their own pages:
- Profound FactCheck. Queries answer engines, uses its own claim-detection model to extract factual statements, checks each against a Knowledge Base you own and scores it accurate or inaccurate. Subjective statements go to Sentiment. Each inaccurate claim is traced to the URL that likely produced it. It names ChatGPT, Perplexity and Gemini as examples, not a full list. A Knowledge Base is required. FactCheck is listed as available to Enterprise customers at no additional cost, and Enterprise pricing is not published. The model's internals, accuracy rates and validation results are not published (Profound FactCheck; Profound help, About FactCheck).
- seoClarity ArcAI Accuracy. Says it monitors AI-generated answers about a brand and fact-checks them against the brand's own trusted content, flagging inaccurate or outdated statements. The page does not name the assistants, describe the method, give monitoring frequency or state pricing. Its statistics are seoClarity's own and not independently validated on the page (seoClarity).
- Semrush AI Visibility Toolkit. Brand Performance reports on the narratives and sentiment around a brand. The knowledge base page does not say whether reports check factual accuracy (Semrush knowledge base). Semrush's blog describes a manual process: test brand, product and category prompts across platforms, check how AI describes you, check product queries separately and trace errors to the third-party pages AI cites (Semrush blog).
- Scrunch. Core is listed at $250 a month billed monthly with a 7-day trial, 15,000 AI responses and 100 fact-checked pages a month, covering 8 assistants. The page does not describe accuracy monitoring directly. "Fact-checked pages per month" is the closest listed item (Scrunch pricing).
Limits. The same Semrush article notes that models do not update in real time, that platform feedback tools have no guaranteed turnaround, and that corrections can take weeks to months. A fact-check tool tells you what to fix and where it came from. It does not fix the answer.
Free route. Write 10 facts a buyer would rely on: what you sell, who it is for, key limits, locations, your own published numbers. Ask each assistant 5 questions that depend on them, monthly, and log mismatches with the cited source. Fix the page you control first, then the third-party page the assistant cited.
#Comparison table: 18 tools by job
All details are from pages read on 21 November 2026. "Not published" means the page I read did not say.
| Job | Tool | What it reports | What the page says about method | Free route on the page | List price read that day |
|---|---|---|---|---|---|
| 1 Research | AlsoAsked | People Also Ask questions | Source not described | None mentioned | $9 a month Basic |
| 1 Research | AnswerThePublic | Questions by keyword | Source not named | Free daily limit, number not stated | US$13.33 a month Starter, billed yearly |
| 1 Research | Google Search Console | Your own queries | Regex filter, RE2 | Price not stated on pages read | Not stated |
| 1 Research | Google Trends | Relative interest | Not published | Not stated | Not stated |
| 1 Research | Profound Prompt Volumes | What users prompt AI engines | Not published | None | Enterprise, custom |
| 2 Formatting | URL Inspection | Whether controls are visible to Googlebot | Google documentation | Part of Search Console | Not stated |
| 2 Formatting | Surfer | Content score, AI tracker | Score not published | Free AI visibility audit | $49 to $299 a month, billed yearly |
| 2 Formatting | Frase | Content audit, AI tracking | Not published | 7-day trial | $39 a month Starter, billed yearly |
| 3 Markup | Rich Results Test | Eligible Google rich results | Google feature requirements | Public tool | Not stated |
| 3 Markup | Schema Markup Validator | General Schema.org validity | Not product-specific | Public tool | Not stated |
| 3 Markup | Screaming Frog SEO Spider | Crawl-wide extraction and structured data checks | XPath, CSS, regex | Free to 500 URLs | 199 pounds a year |
| 4 Tracking | Otterly | Mentions and citations by engine | Credit per prompt per engine | Trial, length not stated | 29 euros a month Lite |
| 4 Tracking | Peec AI | Prompts, 3 of 8 models | Not published | Trial, length not stated | $95 a month Starter |
| 4 Tracking | Ahrefs Brand Radar | Mentions and citations | Index: 581M+ prompts | Not published | From $50 a month |
| 4 Tracking | Semrush AI Visibility Toolkit | Visibility, prompts | Not published | Not published | Not read |
| 4 Tracking | AthenaHQ | Responses by platform | Credit per response | Free Essentials, 300 credits | $295 a month Starter |
| 5 Accuracy | seoClarity ArcAI Accuracy | Fact-checks of AI answers | Not published | Not published | Not published |
| 5 Accuracy | Scrunch | Fact-checked pages per month | Not published | 7-day trial | $250 a month Core |
Profound FactCheck, covered in Group 5, makes 19 products, but it is a feature of Profound, which appears in Group 1, so I count 18 tools. Prices are list prices and the Semrush price was not read.
#How to assemble a working set
- Build the question list. Use Search Console regex filters, Bing grounding queries, and your sales and support questions. Add AlsoAsked or AnswerThePublic only if you need more breadth than your own data gives.
- Fix the pages. Put the answer in the first two sentences, match the heading to the question, and check controls with URL Inspection.
- Validate the markup. Rich Results Test and Schema Markup Validator on one page of each template, then Search Console reports.
- Track. Size by prompts times assistants times days. Start with a manual sheet if you are under about 50 prompts.
- Check accuracy. Ten facts, five questions each, monthly. Buy an accuracy tool only when errors recur and you cannot trace them by hand.
For crawler logs, citation-oriented content tools and earned-mention tools, see the GEO tools guide. For agencies that do this work for you, see GEO agencies compared. For the work behind the measurement, read how to get cited by ChatGPT.
#Methodology
I opened each page listed in Sources on 21 November 2026 using a page-fetch tool and recorded only what the page stated. Prices are list prices on that date, excluding tax, in the currency shown. Prices move often, so check before buying. Page-fetch tools summarise pages, so a figure I could not tie to a visible statement was left out.
Pages that could not be opened or did not carry the fact I needed:
- Bing's help page for the AI Performance report returned a client error, so the two Bing blog posts are used instead.
- Semrush's Keyword Magic Tool page returned a client error and is not used, so no Semrush question research feature is described.
- Ahrefs' Keywords Explorer page did not name its reports or state a price, so it is not used.
- The Schema Markup Validator page did not describe its scope, so its Schema.org announcement and documentation are used.
- Semrush's pricing page was not opened, so no Semrush price appears.
- Google's Performance report help page (7576553) did not state retention beyond the default three-month view, so no retention figure is given.
- Search Console's about page did not state a price, so I do not call it free in my own voice.
The FAQ documentation URL I opened returned Google's updates log, not a standalone feature page, so the FAQ section rests on log entries and on pages that do not list the feature. Redirects followed: Surfer's pricing moved to surfer.ai. Not verified and left out: independent accuracy tests of any tool, any evidence that markup or content scores change an AI answer, and whether Bing or assistants read FAQ markup. Tool counts are mine, and the demand figure is US monthly search volume from data supplied for this site, which I cannot link.
#Sources
- AlsoAsked, pricing. https://alsoasked.com/pricing (read 21 Nov 2026)
- AnswerThePublic, pricing. https://answerthepublic.com/pricing (read 21 Nov 2026)
- Google Search Console Help, Performance report advanced filtering and comparison. https://support.google.com/webmasters/answer/17011165 (read 21 Nov 2026)
- Google Search Console Help, Performance report. https://support.google.com/webmasters/answer/7576553 (read 21 Nov 2026)
- Google Trends. https://trends.google.com/trends/ (read 21 Nov 2026)
- Profound, pricing. https://www.tryprofound.com/pricing (read 21 Nov 2026)
- Bing Webmaster Blog, "Introducing AI Performance in Bing Webmaster Tools Public Preview", February 2026. https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview (read 21 Nov 2026)
- Google Search Central, featured snippets. https://developers.google.com/search/docs/appearance/featured-snippets (read 21 Nov 2026)
- Google Search Central, AI features and your website. https://developers.google.com/search/docs/appearance/ai-features (read 21 Nov 2026)
- Surfer, pricing. https://surfer.ai/pricing/ (read 21 Nov 2026)
- Frase, pricing. https://www.frase.io/pricing (read 21 Nov 2026)
- Screaming Frog, SEO Spider. https://www.screamingfrog.co.uk/seo-spider/ (read 21 Nov 2026)
- Google Search Central, structured data policies. https://developers.google.com/search/docs/appearance/structured-data/sd-policies (read 21 Nov 2026)
- Google Search Central, introduction to structured data. https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data (read 21 Nov 2026)
- Google Search Console, Rich Results Test. https://search.google.com/test/rich-results (read 21 Nov 2026)
- Schema.org blog, "Announcing Schema Markup Validator". https://blog.schema.org/2021/05/21/announcing-schema-markup-validator-validator-schema-org-beta/ (read 21 Nov 2026)
- Schema.org, validator documentation. https://schema.org/docs/validator.html (read 21 Nov 2026)
- Google Search Central, FAQ structured data URL (returned the updates log). https://developers.google.com/search/docs/appearance/structured-data/faqpage (read 21 Nov 2026)
- Google Search Central, documentation updates. https://developers.google.com/search/updates (read 21 Nov 2026)
- Google Search Central, structured data gallery. https://developers.google.com/search/docs/appearance/structured-data/search-gallery (read 21 Nov 2026)
- Google Search Central, QAPage structured data. https://developers.google.com/search/docs/appearance/structured-data/qapage (read 21 Nov 2026)
- Otterly, pricing. https://otterly.ai/pricing (read 21 Nov 2026)
- Peec AI, pricing. https://peec.ai/pricing (read 21 Nov 2026)
- Ahrefs, Brand Radar. https://ahrefs.com/brand-radar (read 21 Nov 2026)
- Semrush, AI Visibility Toolkit (knowledge base). https://www.semrush.com/kb/1493-ai-visibility-toolkit (read 21 Nov 2026)
- AthenaHQ, pricing. https://www.athenahq.ai/pricing (read 21 Nov 2026)
- Semrush, "Fix AI brand misinformation". https://www.semrush.com/blog/fix-ai-brand-misinformation (read 21 Nov 2026)
- Profound, FactCheck. https://www.tryprofound.com/features/answer-engine-insights/factcheck (read 21 Nov 2026)
- Profound help, About FactCheck. https://help.tryprofound.com/articles/5793584301-about-factcheck (read 21 Nov 2026)
- seoClarity, AI search accuracy protection. https://www.seoclarity.net/ai-seo/ai-search-accuracy-protection (read 21 Nov 2026)
- Scrunch, pricing. https://scrunch.com/pricing (read 21 Nov 2026)
- Shruti Kaushik, free AI visibility check (own service, not compared above). https://shrutikaushik.com/ai-visibility-check