SEO Agents
See whether AI search and answer engines can discover, understand and cite your website — and what to fix first.
Paste a URL and the agent examines the signals that determine whether AI-powered search can find your pages, parse what you do, and quote you accurately. You get a scorecard across discoverability, machine readability and citability, plus a prioritized list of changes to review with your team.
Try an example
Search is no longer ten blue links. Buyers increasingly ask ChatGPT, Perplexity, Gemini or Google's AI Overviews for recommendations, and the answer they get is assembled from whichever sources the engine could find, read and trust. The AI Visibility Audit measures how your website performs against that new pipeline — before you assume your rankings from classic search will carry over.
The audit is for founders who suspect competitors are being recommended instead of them, SEO teams extending their practice into generative engine optimization, content leads whose carefully written pages never seem to get quoted, and agencies that need a concrete way to explain AI search to clients.
You give the agent a URL. It examines the things answer engines depend on: whether AI crawlers can access your pages at all, whether structured data identifies your organization and products unambiguously, whether your content states clear, self-contained answers that can be quoted, whether headings and page structure make extraction easy, and whether basic machine-readable signals — sitemaps, clean HTML, consistent entity naming — are in place.
The output is a scorecard grouped by dimension, with each finding explained in plain language and ranked by likely impact. This is deliberately not a generic 'GEO score'. A finding like 'your pricing page cannot be crawled by common AI bots' or 'your product is described three different ways across your site' is actionable; a number without reasons is not.
Because the audit runs inside Dual7, the follow-through is unusually short. Many fixes — adding structured data, restructuring headings, clarifying entity descriptions — can be made directly in your Dual7 project and republished to your domain. Larger content programs can be planned from the findings and routed through the governed pipeline for review.
If AI crawlers are blocked by your robots rules or CDN configuration, nothing else matters. The audit surfaces access problems first.
Discoverability, machine readability, entity clarity and citability are scored separately, so you can see which layer is actually failing.
Each item explains what the agent observed, why answer engines care about it, and what a fix would look like — no unexplained scores.
The audit flags places where your company, product or people are described inconsistently, which makes it harder for models to form a confident picture of you.
Pages are assessed for quotable, self-contained statements — the kind of passage an answer engine can lift — not merely for technical accessibility.
Fixes are ordered so access and entity issues come before polish, giving you a sequence to work through instead of a wall of observations.
Re-run the audit after you ship fixes to confirm each finding is resolved and catch anything the changes introduced.
Point the agent at your homepage, a key landing page, or any URL you want assessed for AI search visibility.
Crawler access, structured data, entity clarity, content structure and quotability are checked against how answer engines consume pages.
Read the findings by dimension, each with an explanation and a suggested fix, ranked by likely impact.
Apply changes in your project, publish to your domain, and run the audit again to verify the findings are resolved.
Checks whether common AI and answer-engine crawlers can reach your pages, including robots directives, meta rules and obvious CDN or firewall blocks.
Examines existing JSON-LD for Organization, Product, FAQ and article markup, and identifies missing or contradictory properties that blur what you are.
Looks at how consistently your brand, products and key people are named and described across the page — the raw material models use to understand who you are.
Assesses whether pages contain direct, self-contained answers to the questions your buyers ask — definitions, comparisons, pricing facts — in a form that can be quoted accurately.
Evaluates whether your heading hierarchy and page layout make the main points extractable, or bury them in unstructured prose and markup.
Findings are compiled into an ordered list — access first, then entity and structure, then content polish — so remediation starts where it matters.
A per-dimension readout covering discoverability, machine readability, entity clarity and citability for the audited URL.
Findings ordered by likely impact, each with the observation, the reason it matters and a suggested change.
Which schema types are present, which are missing or contradictory, and what to add for your kind of site.
Passages that work well as quotable answers, and sections where the key facts are present but not extractable.
A clear statement of which AI crawlers can and cannot reach the page, and the configuration responsible.
Example output
Input: the homepage of a project-management tool. Findings: common AI crawlers blocked by a CDN rule; no Organization schema; the product described as 'a workspace', 'a platform' and 'a tool' on the same page; pricing only visible behind JavaScript. Fix list: allow crawler access, add Organization and Product markup, settle on one product description, and publish a crawlable pricing summary.
A founder whose product never appears in AI answers runs the audit, finds crawler access blocked at the CDN, and fixes the one issue that was hiding everything else.
An in-house SEO lead adds the AI visibility scorecard to their existing audit routine so classic and AI search are reviewed together.
A content lead uses the citability notes to rewrite key pages with direct, self-contained answers instead of marketing prose.
An agency runs audits at the start of an engagement and after each round of fixes, giving clients a concrete before-and-after view.
A developer-tools company audits its docs so answer engines can extract accurate setup instructions instead of guessing from fragments.
A team about to launch a new marketing site audits the staging URL so access and structured data are right from day one.
A company that changed its product name audits key pages to find leftover references to the old name that would confuse entity recognition.
Software companies live and die by recommendations; the audit shows whether answer engines can even describe the product correctly.
Stores check whether product and category pages are readable and quotable when shoppers ask AI assistants what to buy.
Firms verify that their expertise pages give answer engines clear, citable statements about what they do and for whom.
Editorial sites check crawler access and article markup so their reporting can be surfaced and attributed correctly.
Clinics make sure service information is machine-readable and accurately stated, since answers in this space are held to a higher bar.
Course providers and schools audit program pages so prospective students get correct facts when they ask AI for options.
You provide
https://our-saas-product.com
The agent
The agent checks crawler access, structured data, entity naming and content structure across the homepage and linked key pages.
You get
A scorecard showing good content structure but blocked AI crawlers and missing Organization schema, with a two-item fix list to start.
You provide
Audit our docs site — answer engines keep misquoting our setup steps
The agent
The agent examines documentation pages for extractable, ordered instructions and checks whether headings and code blocks are machine-readable.
You get
Citability notes showing steps buried in tabbed JavaScript components, with a recommendation to render core instructions as plain HTML.
You provide
https://shop.example.com — are our product pages visible to AI shopping assistants?
The agent
The agent reviews product page markup for Product schema, consistent naming and crawlable pricing and availability.
You get
A structured data gap report: no Product markup, prices rendered only by script, and a suggested schema block to review.
You provide
We rebranded last year — find anything still describing the old product
The agent
The agent scans key pages for inconsistent entity naming, old product names and contradictory descriptions.
You get
An entity clarity report listing six pages where the old name survives, including one in the schema markup itself.
You provide
https://agency-client-site.com — pre-engagement baseline
The agent
The agent runs a full audit and compiles findings into a prioritized list an account manager can walk a client through.
You get
A baseline scorecard and fix sequence the agency uses to scope the first month of work and re-audit against later.
You provide
Check our new landing page before we launch it
The agent
The agent audits the staging URL for access, markup and quotable content before any promotion starts.
You get
A short findings list — missing FAQ schema, a headline that says nothing concrete — fixed before the page ever goes live.
| Aspect | With the agent | Manual process |
|---|---|---|
| Starting the audit | Paste a URL; the agent reads the live page directly | Check robots files, view source and test crawlers by hand |
| Crawler access testing | Common AI crawlers checked as a standard step | Requires knowing which bots exist and how to test each |
| Structured data review | Present, missing and contradictory markup compiled for you | Manually validating JSON-LD page by page |
| Entity consistency | Naming and description drift flagged across the page set | Reading every page with the problem in mind and hoping to notice |
| Citability judgment | Quotability assessed per section with notes | Subjective review with no structured rubric |
| Prioritization | Findings ranked by likely impact out of the box | You decide what matters after assembling the raw findings |
| Re-checking after fixes | Re-run the same audit and compare | Repeat the entire manual process |
A chatbot can only discuss a page you paste into the context. The agent fetches the live URL and examines what answer engines would actually receive.
Chat answers vary with the prompt. The audit applies the same checks in the same order, so results are comparable across pages and over time.
The scorecard is structured data, not prose — each item has an observation, a reason and a suggested fix you can assign and track.
After you ship fixes, re-running the audit confirms resolution. Asking a chatbot to re-check means starting a new conversation from scratch.
No amount of content polish helps if AI crawlers cannot reach the page. Work the fix list in the order given.
Pricing, product and comparison pages are what buyers ask about. Audit the pages you most want quoted.
Pick a single clear phrase for your product category and use it everywhere — homepage, schema, about page, docs. Consistency is what lets models describe you confidently.
Where a page addresses a common question, state the answer directly in one or two sentences before the nuance. Self-contained statements are what gets quoted.
Schema should describe what is actually on the page. Markup that overstates your offering can be ignored or treated as spam.
Pricing, feature lists and contact details hidden behind client-side rendering may never reach an answer engine. Render essentials as plain HTML.
Redesigns, replatforms and CDN changes routinely break access or markup. Run the audit after any major release.
It is an assessment of how well your website works with AI-powered search and answer engines: whether their crawlers can access your pages, whether structured data identifies you clearly, and whether your content can be accurately extracted and cited. The output is a scorecard with prioritized fixes.
No, and be skeptical of anyone who promises that. Answer engines make their own sourcing decisions. The audit is designed to remove the technical and structural barriers that keep eligible sites from being considered, and to make your content easier to cite accurately.
A classic SEO audit focuses on rankings in link-based results: keywords, backlinks, indexation. This audit focuses on the answer-engine pipeline — crawler access for AI bots, entity clarity, structured data and whether your content is quotable. The two complement each other; most sites need both.
GEO — generative engine optimization — is the practice of making content visible and citable in AI-generated answers. This audit operationalizes the website side of GEO: the technical access, markup and content structure that answer engines depend on.
Just a public URL. The agent reads the live page the way an outside crawler would. For the most useful picture, audit your homepage and the pages buyers most often ask about, such as pricing and product pages.
It is the first thing to fix. If answer-engine crawlers cannot reach your pages, your content cannot be considered for citation at all. The fix is usually a robots or CDN configuration change, and the report identifies what is responsible.
Yes, and you should. Re-auditing after fixes confirms each finding is resolved and catches anything the changes broke. Many teams run it after every significant release.
Founders and marketers who want to know why competitors get recommended, SEO teams extending into generative engine optimization, content leads improving quotability, and agencies that need a concrete AI search deliverable for clients.
The audit examines your own website — the part you control. Off-site mentions and third-party coverage also influence AI answers, and the report notes where those factors matter, but measuring them is outside what a URL audit can observe.
Yes. The scorecard and fix list are yours to share with your team, hand to a developer or work through yourself. If your site is a Dual7 project, many fixes can be applied and republished in the same place.
Dual7 is in early access, which is free with no credit card required. You can run an audit on your URL and review the full findings before deciding anything.
Dual7 Agents
See whether AI search and answer engines can discover, understand and cite your website — and what to fix first. Start with your own input — the output is an editable project you own.
500+ users in production · Kanan.co moved off Salesforce onto Dual7 in four months · 100% code owned
Dual7 · a product of AlgorithmShift · Build once. Own forever.