How We Diagnose SEO Problems on AI-Built Websites
Our audit separates discovery, crawling, rendering, indexation, ranking, and AI-search visibility. It verifies the live production environment, documents evidence, prioritizes fixes, and defines how implementation will be checked. It does not promise rankings, indexation, or citations.
1. Establish the production truth
Status, redirects, and canonical domain
We request priority URLs, follow redirects, verify final status codes, compare host/protocol variants, and identify canonical-domain conflicts.
Raw HTML and rendered output
We compare the initial server response with the browser-rendered page and record the exact tool, date, route, and limitation. A browser screenshot alone is not crawl evidence.
Crawlable route discovery
We test ordinary HTML links, navigation, route responses, orphan risks, and whether deep URLs work when requested directly.
Sitemaps and robots
We validate availability, declared URLs, indexability consistency, robots directives, and relevant response headers. Permission does not guarantee crawling.
2. Test indexation signals and page ownership
When the owner provides Search Console access, we record URL Inspection and indexing evidence without publishing private screenshots. We review self-canonicals, duplicates, route-level titles/descriptions, heading structure, intent ownership, internal links, and template overlap. One page owns each material intent; near-duplicate expert, consultant, agency, and service variants are consolidated.
3. Review content, performance, and machine-readable context
Content and intent
Does the page answer the query with original evidence, clear ownership, useful detail, and an appropriate conversion path?
Core Web Vitals
We distinguish lab diagnostics from field data and disclose sample limitations. Page experience matters to users and broader systems; a green score is not an automatic ranking boost.
Structured data
We validate syntax, eligibility, duplication, and whether markup matches visible content. Schema can clarify entities and support eligible features but cannot guarantee results or AI citations.
AI-search visibility
We review crawl/index foundations, authorship, entity clarity, original information, corroboration, and a repeatable monitoring protocol. Each answer engine controls retrieval and citation.
4. Prioritize by evidence and risk
| Field | What it records |
|---|---|
| Finding | One observable problem or opportunity. |
| Severity | Impact, affected scope, confidence, and business relevance. |
| Evidence | URL, response/render capture, source, and test date. |
| Fix | Specific implementation, owner, dependency, and rollback risk. |
| Verification | Acceptance test and 14/30/60/90-day measurement where relevant. |
What you receive
- Executive summary and prioritized issue register.
- URL/template-level evidence with reproducible checks.
- Implementation guidance, dependencies, owners, and risk notes.
- Keyword/page ownership and internal-link recommendations where in scope.
- Validation checklist and measurement framework.
What the audit does not guarantee
We do not guarantee a crawl date, indexation, ranking position, traffic, lead volume, rich result, AI mention, citation, or recommendation. Search systems and market conditions remain outside our control. We guarantee only the explicitly agreed deliverable and correction of our own implementation defects within the written service terms.
How implementation is validated
After release we recheck status, canonical, robots, server-delivered and rendered content, links, metadata, schema, forms, layout, sitemap inclusion, and cache behavior. We record what passed, what was not tested, and what requires Search Console or owner evidence.
See the Bolt.new SEO service, Lovable SEO service, and AI-search visibility guide.
Start with production evidence
Share a public URL for initial triage, or request a full audit when route, rendering, or indexation problems require manual investigation.
Request free triageView the audit