SEO Audit Case Study: How One Site Won AI Visibility

See It in Action: An SEO Audit Case Study Breakdown (Step-by-Step)

Introduction: what this SEO audit case study will show you (and who it’s for)

I recently audited a mid-sized US service business that was in a precarious spot: their traffic was steady, but their leads were flatlining. They were ranking, but they weren’t converting, and they were terrified of losing visibility to the new AI-driven search features rolling out daily. This isn’t a theoretical exercise; this is a breakdown of the exact workflow I used, the specific issues I found, and how we prioritized fixes with a limited engineering budget—basically, we had one dev sprint to get the technical side right.

In this post, I will walk you through a real-world SEO audit case study that moves beyond the old “check the meta tags” approach. You will see how modern audits must account for AI Overviews, zero-click behaviors, and Core Web Vitals 2.0 metrics like Interaction to Next Paint (INP). My goal is to give you a replicable process—complete with templates and decision logic—so you can run a meaningful audit that doesn’t just produce a list of problems, but a roadmap for business results.

Here is what we will cover:

• The Setup: The exact tool stack and data sources I used.
• The Findings: How high INP scores and unstructured content were killing conversions.
• The Fixes: The prioritization logic I used to decide what to ship first.
• The Results: Honest before-and-after metrics (including where we failed to move the needle).

Quick definitions (so beginners don’t get lost)

Before we dive into the data, let’s clarify the terms that matter most in 2026.

• SEO Audit: A comprehensive checkup of a website’s technical health, content quality, and authority to identify why it isn’t performing.
• Indexability: Whether search engines can actually find, read, and store your pages in their database.
• Core Web Vitals (CWV): Google’s specific metrics for user experience, focusing on loading speed (LCP), visual stability (CLS), and interactivity (INP).
• Structured Data/Schema: Code you add to your HTML to help search engines understand exactly what your content is (e.g., “this is a recipe,” “this is a FAQ”).
• AI Overviews/Answer Engines: The AI-generated summaries at the top of search results that answer questions directly, often removing the need for a click.

Why SEO audits changed: from “rankings only” to AI visibility + zero-click outcomes

For years, the goal of an SEO audit was simple: push rankings higher to get more clicks. But when I review Search Console data now, I am less surprised when I see impressions rising while clicks remain flat. The search landscape has fundamentally shifted. With AI Overviews now appearing in a significant portion of queries—some industry reports suggest up to 47% —the user often gets their answer without ever visiting a website.

This “zero-click” reality means a modern audit cannot just look at blue links. We have to audit for visibility. Are we appearing in the AI summary? Is our brand being cited as the source? Are we winning the Featured Snippet? If users are getting the answer on Google, we need to ensure we are the ones providing that answer to build brand trust and influence later conversions. In this new environment, an audit that ignores engagement reliability or content structure is effectively obsolete.

What distinguishes an SEO audit for AI visibility versus traditional SEO?

The difference lies in what we optimize for. A traditional audit fixes technical errors to please a crawler. An AI visibility audit ensures content is understandable and useful enough to be synthesized by an engine.

• Measurement: We track citation frequency and snippet ownership, not just rank position.
• Changes: We restructure paragraphs into clear “Q&A” formats and use heavy schema markup rather than just stuffing keywords.
• Wins: Success is defined by brand impressions and assisted conversions, rather than just direct organic clicks.

The SEO audit case study: the business snapshot, goals, and constraints

Let’s look at the subject of this audit. I worked with a US-based B2B service provider in the logistics sector. They had about 400 pages of content, a mix of service pages and a stale blog, running on WordPress. Their primary pain point was clear: “Our traffic looks okay on paper, but nobody is filling out the ‘Request a Quote’ form anymore.”

I scoped this audit to focus on three specific areas, while explicitly deciding not to recommend a full site redesign due to budget constraints. We had to work with the site we had.

The Audit Scope:

• Technical Health: Specifically targeting mobile performance and the new INP metric.
• Content Structure: Evaluating if their content was ready for AI extraction.
• User Journey: Identifying why traffic wasn’t turning into leads.

Success metrics I committed to before starting

To avoid moving the goalposts later, we defined what “good” looked like upfront. I intentionally avoided committing to “total keyword count” because it’s a vanity metric that often misleads stakeholders.

• Primary KPI: Qualified organic leads (form submissions).
• Secondary KPI: Non-branded impression growth (visibility).
• Technical KPI: Passing Core Web Vitals (specifically getting INP under 200ms).
• Visibility KPI: Increase in Featured Snippet/AI citation ownership.

Tools, data sources, and how I set up the audit (so you can replicate it)

I believe in keeping the tool stack simple but deep. You don’t need twelve different subscriptions to do a world-class audit. For this case study, I relied heavily on Google’s own data, supported by a crawler and a specialized AI SEO tool to analyze content depth.

The setup wasn’t without friction. We initially realized their Google Analytics 4 (GA4) property wasn’t filtering out internal traffic, skewing our engagement data, and the Search Console property was set to “URL Prefix” instead of “Domain,” missing data from some subdomains. Once we cleaned that up, here is the stack I used:

Data Source	What it Diagnoses	Common Beginner Mistake
Google Search Console	Rankings, CTR, Indexation issues	Ignoring the “Not Indexed” report reasons.
Screaming Frog	Technical crawl (broken links, metadata)	Crawling with JavaScript rendering off (missing content).
PageSpeed Insights	Core Web Vitals & Real user data	Looking only at “Lab Data” and ignoring “Field Data.”
Kalema	Content relevance & topic gaps	Using an SEO content generator blindly without strategy.

Audit setup checklist (copy/paste)

1. Confirm Indexation: I checked the “Coverage” report in GSC to ensure high-value pages were actually indexed.
2. Export Top Pages: I pulled the top 50 pages by traffic and the top 50 by impressions (but low clicks).
3. Crawl the Site: I ran a full crawl to catch broken links and orphan pages.
4. Capture CWV Baseline: I recorded the LCP and INP scores for the top 10 templates.
5. Map Key Intents: I categorized key pages by “Informational” vs. “Commercial” intent.

Technical findings: indexation, crawl efficiency, and Core Web Vitals (CWV 2.0)

The technical analysis revealed why the site felt “sluggish” despite decent server speeds. It wasn’t the server; it was the code execution. Specifically, the site had a massive Interaction to Next Paint (INP) issue. The home page had a heavy JavaScript slider that delayed browser responsiveness. When users clicked “Get a Quote,” there was a perceptible 400ms lag before anything happened. In the mobile world, that lag feels like a broken button.

We also found indexation bloat. The site was generating thousands of parameter URLs from product filters (e.g., ?color=blue&sort=desc) that were wasting Google’s crawl budget. Here is the technical breakdown we found:

Metric	Target Threshold	What We Saw	The Fix
LCP (Loading)	< 2.5s	3.8s (Mobile)	Compressed hero images & deferred off-screen images.
INP (Interactivity)	< 200ms	410ms (Poor)	Removed heavy third-party chat widget & optimized main-thread JS.
CLS (Stability)	< 0.1	0.25 (Poor)	Added explicit width/height to images to stop layout shifts.

What technical performance metrics are now essential in SEO audits?

In 2026, Core Web Vitals are table stakes. If you aren’t measuring these, you aren’t seeing what Google sees.

• INP (Interaction to Next Paint): Measures responsiveness. High INP = the page feels frozen when you tap.
• Engagement Reliability (ER): An emerging concept regarding how consistently a page functions across devices.
• LCP (Largest Contentful Paint): How fast the main content loads.

High-impact technical fixes I prioritized first

We couldn’t fix everything. I chose these three because they impacted 90% of the site’s traffic:

1. Canonical Tag Cleanup: We pointed all parameter URLs back to the main category pages to consolidate ranking signals.
2. Image Optimization: We installed a plugin to convert images to WebP automatically.
3. Script Cleanup: We removed a heatmap tracking script that nobody had looked at in two years but was blocking the main thread.

Content and on-page findings: what I changed to earn clicks (and citations) in 2026 SERPs

The content audit was where we found the biggest growth opportunity. The client had excellent expertise, but their pages were walls of text. A user (or an AI bot) had to read 800 words just to find the definition of a core industry term. This is fatal for modern SEO. We needed structure.

We used an AI article generator as a drafting assistant to help rewrite introductions and generate schema-friendly summaries, which we then edited heavily for tone. This allowed us to refresh older content much faster than writing from scratch.

The Transformation Example:
Before: A 5-line intro paragraph burying the definition of “logistics compliance.”
After: A bold list item: “Logistics compliance is the process of adhering to…” followed by a bulleted list of key regulations.

Page Type	User Intent	Primary CTA	Elements We Added
Service Page	Commercial / transactional	Request Quote	Pricing tables, Process checklist, Trust badges.
Blog Post	Informational	Newsletter / Read More	“Key Takeaways” box, FAQ Schema, Definition blocks.
Case Study	Investigational	Book Demo	“Results at a glance” bullet list, Client quote.

How I structured pages for AI-generated answers (without sacrificing humans)

I used to write clever, winding introductions. Now, I write for extraction. Here is the checklist we applied to the top 20 pages:

• Definition First: Define the core topic in the first 100 words (Subject + Predicate + Object).
• Scannable Headers: Use H2s and H3s that ask questions users are actually searching for.
• Lists & Tables: AI engines love structured data. If it can be a list, make it a list.

Topic clusters: the content gap that moved visibility the fastest

We found that while they had a “pillar” page for “Freight Forwarding,” they lacked the supporting content to establish authority. We built a cluster around it:

• Center: Freight Forwarding Services (Pillar)
• Spoke 1: Cost of Freight Forwarding in 2025
• Spoke 2: Freight Forwarding vs. 3PL
• Spoke 3: International Customs Checklist

Internal linking between these pages signaled to Google that we covered the entire topic, not just the sales pitch.

AI visibility + multimodal findings: measuring beyond rankings (AI Overviews, video, images, voice)

We stopped tracking just “rankings” and started tracking “pixels occupied.” We wanted to know if we were appearing in the AI Overviews or image packs. We found that for many technical terms, the client was ranking #4, but the AI Overview was taking up the entire top of the screen. We weren’t winning the click because we weren’t part of the answer.

One limitation I always tell clients: AI Overview presence is volatile. I treat it as a trend line, not a daily KPI. If we are cited in 20% of queries this month and 30% next month, we are winning.

What I consider a win in the AI era:

• Cited Source: Our link appears in the AI Overview footnotes.
• Snippet Ownership: We hold the “position zero” box.
• Visual Visibility: Our diagrams appear in image search for process-related queries.

The signals that increased ‘citable’ content blocks

To increase our chances of being cited, we created reusable content block templates for the writers:

• The “Direct Answer” Block: A 40-word bold summary of the answer.
• The “Comparison” Table: X vs Y with clear winner/loser criteria.
• The “Step-by-Step” List: Numbered instructions using active verbs.

SEO audit case study action plan: how I prioritized fixes and shipped improvements

An audit is useless without a roadmap. We had hundreds of issues, but we only had bandwidth for a few dozen fixes. I used an ICE scoring model (Impact, Confidence, Ease) to prioritize. We sequenced technical fixes first (to clear the path for crawlers) and then moved to content updates.

To keep the content updates consistent, we integrated an Automated blog generator into our workflow. This didn’t replace our writers; it gave them a consistency layer, ensuring that every new post already had the correct internal links and schema placeholders before a human editor reviewed it for brand voice.

Issue Category	Specific Fix	Owner	ETA	Expected Impact
Technical	Fix INP on core templates	Dev Team	Week 1	High (Conversion + Rank)
Content	Refresh “Freight” Pillar Page	Content Lead	Week 2	High (Authority)
Schema	Add FAQ Schema to top 20 pages	SEO	Week 2	Medium (Visibility)
Technical	Fix 404 errors on blog	SEO	Week 3	Low (Hygiene)

My prioritization rubric (impact, effort, confidence)

I score everything out of 10. If a task has High Impact (9), High Confidence (9), but Low Ease (2), it might get deprioritized for a “Quick Win” that is Impact (7) and Ease (9). This prevents the team from getting bogged down in massive projects that show no immediate return.

Results and reporting: the before/after metrics I used to prove impact

After 90 days, the results were clear. We didn’t see a massive spike in raw traffic—in fact, clicks were only up slightly. But the business impact was undeniable. The traffic we were getting was engaging. They weren’t bouncing instantly because the page loaded fast (thanks to INP fixes) and they found the answer quickly (thanks to structured content).

What didn’t change: Our total number of keywords ranked remained relatively flat. This is common. We didn’t create hundreds of new pages; we optimized the existing ones. Don’t panic if vanity metrics don’t skyrocket.

Metric	Baseline	90-Day Outcome	Notes
CWV (INP)	410ms (Fail)	140ms (Pass)	Mobile conversion rate improved by 18%.
Organic Leads	22 / month	34 / month	Qualified requests increased significantly.
Impressions	45k / month	62k / month	Strong growth in visibility/awareness.
AI/Snippet Wins	3 owned	11 owned	Captured key definitions in our niche.

How I reported wins when clicks were flat (zero-click reality)

When leadership asked why traffic didn’t double, I showed them the executive summary:

• “We are appearing in 37% more search results (Impressions).”
• “Our brand is answering the customer’s question directly on Google.”
• “The people who click are converting 18% better because the site isn’t slow.”

Common SEO audit pitfalls (and how I avoid them)

I have made plenty of mistakes in past audits. Here are the ones you should avoid:

1. Auditing without goals: If you don’t know what business outcome you want, you’re just making lists of chores.
2. Ignoring the baseline: I used to forget to screenshot the “Before” stats. Without them, you can’t prove your value later.
3. Chasing green scores: Getting a 100/100 on Lighthouse is vanity. Getting a 92/100 and shipping a new landing page is strategy.
4. Fixing low-traffic pages first: Don’t spend a week optimizing a blog post from 2018 that gets zero visitors. Prioritize your money pages.
5. Forgetting internal links: It is the easiest, highest-ROI fix you can make, yet it’s often skipped.

FAQs: what beginners ask me after reading an SEO audit case study

How is an AI visibility audit different from a regular one?
Regular audits focus on technical crawling and keyword placement. AI visibility audits focus on content structure, entity clarity, and citation readiness (formatting content so machines can easily summarize it).

What metrics actually matter for performance now?
Focus on INP (Interaction to Next Paint) for interactivity, LCP for load speed, and CLS for stability. Emerging metrics like Engagement Reliability are also becoming important for understanding consistent user experiences.

How do I structure content for AI answers?
Use a “reverse pyramid” style. Start with the direct definition or answer. Follow with a bulleted list of details. Use clear, descriptive headings. Add FAQ schema.

Why are case study examples effective in SEO audit content?

Case studies provide E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). They prove you haven’t just read the documentation—you’ve actually done the work. Sharing a screenshot of a messy crawl or a change log builds more trust than a thousand words of theory.

Conclusion: my 3-point recap + next actions you can take this week

If you take nothing else away from this case study, remember this:

• Prioritize Impact: Don’t fix everything. Fix the things that block users or crawlers on your most important pages.
• Structure for AI: Format your content to be cited, not just read.
• Measure Business Wins: Track leads and visibility, not just clicks.

Your next actions for this week:

1. Check your Core Web Vitals in GSC—specifically INP.
2. Identify your top 5 pages and rewrite the intros to be “definition-first.”
3. Run a crawl to find and fix your top 3 technical errors.
4. Create a simple 30-day roadmap. Start with one template and one page—you’ll get there.