AI SEO Visibility Case Study Framework for Law Firms

Arslan SEO Insights tells law firms that a real AI visibility case study has to show three things together: what the site looked like before, exactly what structural changes were made, and honest evidence of a citation or mention afterward, connected with a clear explanation of why that change caused that result.

A screenshot of a ChatGPT answer mentioning a firm's name is not proof of anything on its own. Without the starting point and the specific changes made in between, it is just a snapshot with no way to judge whether the work actually caused it.

Why AI Visibility Is Different From a Ranking

A traditional SEO case study leans on a rank tracker. A keyword sits at position 14, work gets done, and three months later it sits at position 6. That is a clean, verifiable, repeatable measurement.

AI visibility does not work that way. When someone asks ChatGPT, Google's AI Overviews, or Perplexity a question like "who handles mesothelioma claims in Texas," the answer generated is not fixed.

It can change from one day to the next, one user to the next, and one phrasing of the question to the next, based on how the AI model retrieves and weighs information at that moment.

There is no single stable position to track the way there is with a keyword ranking.

This means an AI visibility case study has to be built differently than a ranking case study.

Instead of a single number moving up a chart, it needs to document a pattern: the same or similar question asked multiple times across a period, showing the firm getting mentioned or cited with reasonable consistency, not a single lucky screenshot from one afternoon.

What a Credible AI Visibility Case Study Actually Shows

The starting point, documented honestly. Before any changes, was the firm showing up at all when a prospective client asked an AI tool a relevant question? This needs to be checked directly, not assumed.

Ask the actual questions a prospective client would ask, across a few different AI tools, and record what comes back, including if the answer is nothing or if a competitor gets mentioned instead.

The specific structural changes made. This is the part most AI visibility claims skip, and it is the part that actually matters. Vague language like "we optimized the site for AI search" says nothing.

A real case study names the actual changes. One example is restructuring a practice area page so the direct answer to the core question appears in the first paragraph, instead of being buried after three sections of firm background.

Another is adding clear entity information about the firm and its attorneys through structured author bios.

Others include improving schema markup so structured data clearly identifies the firm, its practice areas, and its location, or adding a genuine FAQ section that answers the exact kind of question a person is likely to type into an AI tool.

The result, shown without exaggeration. This means actual documented instances of the firm being mentioned or cited in an AI-generated answer after the changes, ideally checked more than once over a period of weeks, not a single screenshot presented as if it proves a permanent outcome.

Why it matters for the firm, not just that it happened. A citation in an AI answer is only useful if it can plausibly lead to a call or a form submission.

A real case study connects the citation to what it likely means for the firm's intake, while being honest that AI visibility is still a newer and less measurable channel than traditional search traffic.

Why This Is Harder to Prove Than a Keyword Ranking

Rank trackers pull data from Google on a consistent, repeatable basis using standardized methodology. There is no equivalent standardized tool yet for tracking AI citations at scale with the same reliability.

Checking AI visibility today generally means manually asking the same or similar questions across different tools on a recurring basis and recording what comes back, since automated tracking for this is still new and inconsistent across different AI platforms.

This also means a single before-and-after screenshot proves very little by itself.

AI answers can differ based on how a question is phrased, what other content has recently been published on the topic, and even random variation in how the model generates a response at that moment.

A credible case study acknowledges this variability directly, rather than treating one favorable screenshot as if it represents a stable, permanent citation the way a stable page-one Google ranking might.

What Makes This Kind of Case Study Credible Versus Misleading

The difference between a credible AI visibility case study and a misleading one usually comes down to specificity and honesty about uncertainty.

A misleading version says something like "we got our client featured in ChatGPT answers" with a single screenshot and no explanation of what was actually done to the site, or how many times the citation was checked and confirmed.

It implies a guaranteed, permanent outcome from a one-time event.

A credible version names the actual structural changes made to the page, in enough detail that another SEO person could look at the before and after and understand exactly what changed.

It shows the citation checked more than once, across a reasonable span of time, and it says plainly if results were inconsistent, since inconsistency is the honest reality of this kind of work right now.

It also does not claim the AI visibility work alone explains every downstream result.

If a firm's overall case inquiries went up during the same period as an AI visibility project, a credible case study separates what can reasonably be attributed to that specific work from what might be coming from other marketing efforts happening at the same time.

Since this kind of work has been applied to a real, active law firm client rather than a broad roster of clients, any case study built from this experience reflects that one specific engagement in detail, not a generalized claim about results across many firms.

The Actual Framework, Step by Step

Step one: document the baseline. Before making any changes, write down a list of realistic questions a prospective client in the firm's target practice area and location would type into an AI tool.

Ask each question across two or three different AI tools, and record exactly what comes back, word for word, including if the firm is not mentioned at all or if a specific competitor is mentioned instead.

Step two: audit the page for AI readability. Look at whether the page answers its core question directly and early, in plain language, or whether the real answer is buried under firm history, generic disclaimers, or marketing language before getting to the point.

AI tools tend to pull from content that states a clear, direct answer in a short, well-structured passage.

Step three: make the structural changes and document them.

This might include rewriting the opening section of a page to directly answer its core question, adding a genuine FAQ section built around real search questions, improving schema markup with LegalService, Attorney, and FAQ structured data, and strengthening entity signals like clear attorney bios with real credentials and case experience.

Every change made should be written down with a date, so the case study has an accurate record of exactly what happened and when.

Step four: recheck the same questions on a recurring basis. Ask the same baseline questions again a few weeks after the changes, and continue rechecking periodically, not just once. Record every result, favorable or not.

Step five: build the case study from the full record. Present the before state, the specific changes with dates, and the pattern of results over the full checking period, including any inconsistency.

Explain plainly what this likely means for the firm's visibility and intake, without overstating certainty the data does not support.

Common Mistakes to Avoid

Presenting a single screenshot as definitive proof is the most common mistake, since one favorable AI response does not establish a pattern.

Vague descriptions of the work done, like claiming a page was "optimized for AI" without saying what that actually meant in practice, undermine credibility with anyone who understands this space.

Ignoring or hiding inconsistent results is another mistake. If the same question sometimes returns the firm and sometimes does not, that is useful and honest information, not something to leave out of the case study.

Attributing every positive change in the firm's overall traffic or inquiries to AI visibility work alone, when other SEO or marketing work happened during the same period, overstates what can actually be claimed.

Implying the result is permanent or guaranteed is a mistake with real consequences for a law firm specifically, since state bar advertising rules generally restrict this kind of guarantee, and AI visibility is genuinely one of the least stable, most changeable parts of search right now.

How This Fits Into a Broader SEO Strategy

AI visibility work should not replace the fundamentals of a solid law firm SEO strategy. It builds on the same foundation: accurate, genuinely useful content, clear entity signals about the firm and its attorneys, and technical basics like proper schema markup.

A firm with thin, generic practice area pages is unlikely to see meaningful AI visibility improvement no matter how much specific "AI optimization" work is layered on top, since AI tools tend to pull from content that already demonstrates real depth and clarity.

This is part of why AI visibility work is best treated as an extension of ongoing SEO work rather than a separate, standalone service.

The same page improvements that help a page rank better in traditional Google search, like clearer structure, genuine depth, and accurate legal information, also tend to be exactly what helps that page get pulled into an AI-generated answer.

Setting Realistic Expectations

A firm considering this kind of work should expect a genuine testing and documentation period before drawing conclusions, not an instant, guaranteed citation.

Because AI answers vary by tool, by phrasing, and over time, movement here should be judged over a period of months, similar to how traditional rankings need time to stabilize, generally with initial signals appearing around 90 days and a clearer, more stable pattern forming over 6 to 12 months.

Anyone promising a fast, guaranteed AI citation is not describing how this technology actually behaves today.

Tools and Methods for Checking AI Visibility Today

Since there is no single standardized tracker for AI citations the way there is for keyword rankings, checking AI visibility today mostly comes down to a disciplined manual process, done consistently.

Keep a running spreadsheet of the exact questions being tracked, phrased the way a real prospective client would type them, not the way an SEO person would phrase a keyword.

A real question sounds like "can I still sue if I signed a settlement release without a lawyer" rather than a short keyword phrase.

Check each question across the AI tools that matter most for the firm's audience, which today generally means Google's AI Overviews, ChatGPT, and Perplexity.

Record the date, the exact wording of the question, and the full text of the answer, including whether the firm was mentioned, whether a competitor was mentioned instead, and whether any source link was cited alongside the answer.

Repeat this on a set schedule, such as every two to four weeks, rather than only checking once.

A pattern across several checks is far more meaningful than any single check, since it shows whether a citation is a one-time fluke or a repeatable outcome tied to the structural changes made.

Where a firm uses rank tracking software that has started adding AI Overview tracking features, that data can supplement the manual checks, but it should not fully replace them yet, since coverage and accuracy of these newer tools still varies.

A Realistic Example Walkthrough

Consider a mass tort page covering a specific pending litigation. Before any changes, someone asks ChatGPT and Perplexity a direct question like "what is the average settlement range for this litigation" and "how do I know if I qualify to file a claim."

The firm is not mentioned in either answer. A couple of large national firms and a legal news site are cited instead.

Looking at the firm's existing page, the reason becomes clear. The page opens with two paragraphs about the firm's history and awards before ever addressing eligibility or the litigation's current status.

There is no FAQ section. Schema markup on the page is generic, with no LegalService or FAQ structured data present.

The structural changes made would include rewriting the opening of the page to state plainly, in the first two or three sentences, what the litigation is about and who may qualify to file a claim.

A genuine FAQ section gets added addressing the exact two questions asked above, along with a few related ones like what documentation is typically needed and how long the claims process tends to take.

FAQ schema and LegalService schema get added and correctly filled out. The attorney handling this practice area gets a stronger bio section with real, specific experience relevant to this type of case.

Four weeks later, the same two questions get asked again across the same tools. Maybe one tool now includes the firm in its answer and one still does not.

That is a realistic, honest outcome, and it is exactly why continued rechecking over time, not a single test, is what actually shows whether the changes are working.

How to Present This to a Skeptical Firm Owner

Many law firm owners have already seen inflated SEO claims before, and AI visibility is an area where inflated claims are especially easy to make since there is no simple, verifiable number to point to the way there is with a keyword ranking.

The right way to present this work to a skeptical owner is to lead with the documentation process itself, not with a promised outcome.

Show the actual baseline check, including any unfavorable results where the firm was not mentioned or a competitor was. Explain the specific, concrete changes being proposed for the site, in plain language a non-SEO person can follow.

Set expectations that results will be checked and reported on an ongoing basis, with real answers shown each time, not just the favorable ones.

This approach takes longer to explain than a simple promise, but it holds up under scrutiny in a way that a single flattering screenshot never will, and it matches how a law firm owner is trained to think anyway, since they deal in evidence and documentation every day in their own casework.

Next Step

If you want to understand what AI visibility work would actually involve for your firm's specific practice areas and markets, get a free audit, or see AI SEO for Law Firms for how this fits into a full strategy.