On-page SEO
Whether important pages are crawlable, indexable, internally linked, and written around a clear buyer/search intent.
- •Titles and meta descriptions
- •H1/H2 structure
- •Canonical URLs
- •Internal links
- •Indexable page depth
Audit Methodology
How LeadByAI Scores Website Visibility
Our website audits score the signals that determine whether a site can be found, understood, cited, trusted, and converted — by humans, search engines, and answer engines.
Seven-area scorecard
What gets scored
Each dimension is scored from live evidence. We do not treat one technical file, one schema block, or one AI-search trick as a replacement for a useful, crawlable, internally linked website.
AEO readiness
Whether the site gives answer engines direct, citeable, plain-language answers and whether the brand is actually surfaced and cited in observed AI/search samples.
- •Question-and-answer sections
- •Visible source text
- •Direct AI citation samples
- •Search Console generative-AI reporting when available
- •llms.txt as an optional resource index
GEO + local/entity visibility
Whether search systems can understand who the business is, where it truthfully operates, who it serves, and whether independent profiles, citations, reviews, and discovery results corroborate those claims.
- •Organization/entity schema
- •Truthful service-area signals
- •NAP and contact consistency
- •Verified official profiles and reviews
- •Observed local discovery and citation evidence
Content + authority
Whether the site proves expertise with substantive, non-commodity evidence: methods, implementation detail, original observations, limitations, and dated source-backed analysis.
- •Depth of service pages
- •Original research or implementation examples
- •Methodology and limitations
- •Named authorship and review
- •Independent corroboration where available
Technical + crawl health
Whether bots, users, and emerging browser agents can fetch, render, understand, and navigate the site without avoidable friction.
- •Sitemap and robots behavior
- •Redirects and status codes
- •Schema parseability
- •Mobile rendering and performance
- •Lighthouse Agentic Browsing as an informational QA gate
ADA / accessibility risk
Whether forms, navigation, contrast, labels, tap targets, reflow, and content structure meet practical WCAG 2.2 expectations; automation is kept separate from manual and assistive-technology evidence.
- •Automated axe checks
- •Keyboard and focus paths
- •320px reflow and zoom
- •Representative manual review
- •Assistive-technology evidence when available
Overall visibility health
A weighted judgment of how findable, understandable, trustworthy, and conversion-ready the site is today.
- •All six underlying categories
- •Severity of blockers
- •Speed of in-control fixes
- •Competitive baseline where available
Evidence discipline
How the audit is run
01
Collect live evidence
We fetch the public site, robots.txt, sitemap files, representative pages, rendered screenshots, schema blocks, metadata, and crawl status evidence from the current live web.
02
Pin the model and evidence date
Before scoring, we record the model ID and hash, current tool versions, crawl timestamp, and evidence availability. A higher target never changes the scoring rules mid-run.
03
Score observed and unknown signals separately
Each category is scored from fresh evidence. Missing Search Console, review, citation, field-performance, direct-AI, manual-accessibility, or assistive-technology data stays unavailable instead of receiving assumed credit.
04
Map fixes by control and impact
We separate fast in-control fixes from deeper content, design, platform, and authority work so teams know what can move first.
05
Verify before reporting
For implemented changes, we build or serve the site, check critical URLs, parse schema, verify robots/llms/sitemap bodies, and capture visual evidence.
AI-search standard
What we optimize for now
Durable fundamentals
Crawlable pages, useful visible answers, internal links, accurate structured data, entity proof, local signals, performance, and accessibility still do the heavy lifting.
Observed outcomes
Direct citations, Search Console generative-AI reporting, local discovery, reviews, profiles, and independent mentions are scored as outcomes. Technical readiness alone does not receive those points.
Assistant-friendly context
We add curated resource indexes like llms.txt, but treat them as optional supporting context—not a Google ranking lever or substitute for real pages.
Emerging agent readiness
We run Chrome Lighthouse Agentic Browsing for semantic markup, labels, focus, and agent-ready interactions. Because Chrome marks the category informational and unbenchmarked, it remains a QA gate rather than an invented score weight.
Common questions
Do AI Overviews require special AI-only schema?
No. Google Search Central states there is no special AI-only schema or machine-readable file required for AI Overviews or AI Mode. We still use accurate Schema.org markup where it matches visible content because it helps search systems understand the page.
Why include llms.txt if Google does not require it?
LeadByAI treats llms.txt as a curated resource index for inference-time assistants. It supports discovery and context, but it does not replace crawlable pages, internal links, robots.txt, sitemaps, canonical tags, or visible source text.
How do you measure generative-AI visibility?
We combine direct citation samples, standard search discovery, and—when the client can provide it—Google Search Console generative-AI performance reporting. We keep observed outcomes separate from owned-site readiness so technical preparation cannot be mislabeled as market visibility.
Does Lighthouse Agentic Browsing affect the official score?
Not yet. Chrome describes the category as experimental and informational. We run it as a hard QA gate for semantic and agent-ready regressions, but we do not invent an official weight until a benchmarked industry standard exists.
Is the ADA score a legal compliance certification?
No. The ADA/accessibility score flags practical user-friction and risk signals. It is not legal advice and is not a formal WCAG or ADA certification.