How to Read a Deep Audit Report: All 9 Sections Explained

A Deep Audit returns a lot more than a number. One call gives you an LAI score, three sub-scores, nine section scores, a coverage figure, and a ranked list of findings pulled from 14,000+ individual checks. If you've never read one, the output can look dense. This guide walks through every part so you can go from "here's a report" to "here's what I'm fixing first" in about five minutes.

We'll use the real report from our personal-injury law case study as the worked example.

The headline: LAI score and confidence

LAI Score: 3.99/5  (good)   confidence 67.41%
AI 4.1/5   SEO 2.96/5   Local n/a
  • LAI score (0–5) is the overall grade. Bands: <2 poor, 2–3 fair, 3–4 good, 4+ excellent. Most real sites land 3–4 — "good, not great."
  • Confidence reflects how much of the site the engine could evaluate on this run. Lower confidence usually means JS-heavy rendering or blocked resources, not that your site is bad.
  • Sub-scores split the LAI into AI (answer-engine readiness), SEO (classic search), and Local (shown only when the site is a local business). The gap between AI and SEO is the single most useful signal — in our example, 4.1 on AI vs 2.96 on SEO immediately told us where the work was.

The nine sections

Each section is scored 0–5. This is your triage map — sort ascending and start at the bottom.

# Section What it grades
S1 Technical Foundation & Rendering Does the page render, respond, and serve valid markup
S2 Crawlability & Site Architecture Can bots reach and understand your URL structure
S3 Content Quality & Entity Semantics Depth, clarity, and machine-readable meaning of content
S4 Core Web Vitals & Performance Load speed, layout stability, responsiveness
S5 Technical SEO & Index Control Canonicals, robots, sitemaps, indexation
S6 Schema, Local, Trust & E-E-A-T Structured data and trust signals
S7 Security & Privacy HTTPS, headers, privacy posture
S8 Multimodal Content & Media Image alt text, video markup, media SEO
S9 Off-Site Signals & Reputation External signals and reputation

In the example, S6 (Schema/Trust) at 2.77 was the clear outlier — so that's where we started.

Coverage: how complete the run was

Tasks scored: 6027  (passed 2594, warned 1809)
Coverage: 41.9% (100/158 subsections, 125 AI checks run)

Coverage tells you what fraction of the possible checks actually ran. AI-powered checks only fire when the relevant content exists to evaluate, so a lower coverage number is normal and doesn't lower your score — it just means fewer checks applied to your page.

Findings: the actual to-do list

Findings are ranked by severity and tagged with a section + task ID + evidence:

[critical] S6 6.1.01: No Organization schema found
[critical] S2 2.7.02: Text to HTML ratio: 4%

The task ID (6.1.01) is stable, so you can map a fix directly to the check it clears and re-audit to confirm.

One important caveat: the engine weights some schema/local findings critical for types that may not apply to your business — Product schema on a law firm, Event on a SaaS tool. Filter those out first and focus on findings relevant to what your site actually is. Our schema-by-business-type checklist shows which types matter for each vertical.

Turning the report into a plan

  1. Read the AI/SEO gap — it tells you whether your problem is answer-engine readiness or classic search.
  2. Sort sections ascending — the two lowest are 80% of your ROI.
  3. Filter findings to what applies to your business model.
  4. Fix in task-ID order, re-audit, and watch the section score move.

That loop — audit, read, fix, re-audit — is exactly how we lifted our own homepage's schema section into the 4s. The report isn't a grade to feel bad about; it's a prioritized backlog.


Start reading your own reports. Subscribe to a plan — Pro ($39/mo) and Ultra ($99/mo) include unlimited Deep Audits, so you can audit, fix, and re-audit every site you own as often as you want. See plans and pricing →