AI Cites Directories 7x More Than Firms

A buyer types “best mesothelioma law firm in Phoenix” into an AI. The answer comes back in seconds. It looks authoritative. It cites sources. And in roughly 93% of cases we’ve sampled so far, the cited source isn’t the firm’s own website. It’s a directory, an aggregator, or a news article that happened to quote the firm. The firm’s name appears in the answer. Its page doesn’t.

That’s the gap this study measures.

The 7.2x Gap

Our directional data points to a 7.2x citation advantage for aggregator and earned-media pages over a firm’s own domain when AI answers PI and mass-tort queries. Put plainly: when an AI cites a source, it’s seven times more likely to pull from a Martindale page or a newspaper article than from the firm’s website.

This number is illustrative based on early sampling. The full methodology is in progress. But the direction is consistent across every query category we’ve tested.

Here’s what we’re seeing in the data:

• Firm-owned pages appear as cited sources in roughly 6% of AI answers to intent-heavy legal queries.
• Third-party aggregators and directories account for roughly 43% of cited sources in the same query set.
• Earned media placements (news coverage, quoted expert appearances, legal journalism) make up most of the remaining citations.

The math is unsurprising once you understand how AI models weight authority. What’s surprising is how few firms have done anything about it.

Why AI Ignores Your Website

AI citation engines don’t reward keyword density or domain age. They reward epistemic authority: the sense that a source is being talked about, quoted, and referenced by others.

A directory listing has hundreds of firm profiles linked to it. A legal news article that quotes your managing partner has other publications linking to it. Your “Areas of Practice” page has… your own internal links.

The model treats external consensus as a trust signal. A firm’s own site, no matter how well-built, reads as self-referential. The aggregator reads as a community verdict.

There’s a compounding effect here. Once a directory gets cited in AI answers, people click it. More links form. The citation authority deepens. Firms that rely on owned media are running faster to stay in place while aggregators compound.

What the Data Will Confirm

The full study will run across approximately 8,400 PI and mass-tort queries across 150+ metros, building a citation graph of roughly 60,000 individual source attributions. We’ll map each citation to one of four categories: firm-owned, directory or aggregator, earned media (primary source), or earned media (secondary quote). The 7.2x figure will be confirmed, adjusted, or refined based on that full dataset.

We’ll also run a consumer panel to test whether AI-generated answers influence which firm a prospective client contacts first. The early directional answer is yes, significantly.

The Firms That Flip the Ratio

Some firms beat the pattern. Not by gaming directories. By publishing original data.

When a firm releases a study, a market analysis, or proprietary intake data, that content gets cited. Legal journalists reference it. Directories link to it. Other law firm sites quote it. The firm’s domain becomes the primary source instead of the terminal destination.

We’ve tracked this across a small comparison set so far. Firms that publish original research appear in AI citations at roughly 3x the rate of firms with comparable domain authority but only practice-area content. The citation type also shifts: instead of appearing as a named entity in someone else’s source, the firm’s own page gets pulled.

That’s the mechanism. Be the source, not the subject.

Why This Matters to a CMO or CFO

AI search isn’t a future consideration. It’s where a growing segment of buyers start their search right now. Among younger plaintiffs and among buyers in complex litigation (mass tort, product liability, class action), the share using AI to pre-qualify firms before making contact is climbing every quarter.

If your firm isn’t being cited, it’s being screened out before the buyer reaches your site. Traffic analytics won’t show you this. Your form submission volume will.

The firms that move on this early capture citation share while it’s still low-competition. The firms that wait will spend two to three years trying to claw back ground that compounded against them.

What to Watch For

We’ll publish the full dataset when the methodology completes. In the meantime, three things to track internally:

Your AI presence rate. Run your target queries through the major AI tools. Count how often your firm appears vs. how often a directory appears in your place. The ratio is your baseline.

Your citation source. When your firm does appear in an AI answer, what’s being cited? If it’s always a directory or a news mention, your owned content isn’t pulling weight.

Your original data footprint. How many pieces of content on your site make claims that only your firm can make? Original intake data, outcome data, original research? That’s the asset class that gets cited. Everything else is inventory.

---

This study is in preview. The 7.2x figure is directional based on early sampling across PI and mass-tort query categories. Full methodology, N sizes, and confirmed figures publish when the complete dataset closes. Questions on the methodology go to the research team.