Even with the Warmest Referrals, Prospective Clients Are Googling You

The Before Times

For decades, a referral meant the prospective client immediately picked up the phone after a colleague recommended you. Because the referrer vouched for you, there was no need to vet you; the recommendation was proof enough. Those days are long gone.

As the internet expanded, the lay public became more informed. Online searches gave prospects the means to make more informed choices, no matter the strength of the referral.

It went like this:

1. A trusted relationship passed along your name to a prospective client
2. The prospect conducted a Google search that served up your website bio, LinkedIn profile, directory listings, blog posts, etc.
3. After reviewing those online resources, the prospect called or emailed you
4. The attorney-client relationship began

That handoff still happens, but Step 2 is dramatically different.

Agentic AI Is the New Black

The basic Google search is now either secondary or nonexistent. Instead, the prospect is asking ChatGPT, Claude, Gemini, or Perplexity about you before deciding whether to reach out. And if they run a Google search at all, Google is now generating an AI Overview (AIO)—that synthesized answer block that appears above traditional search results—to render a judgment about you and your firm before the prospect clicks any links. Most of them trust it.

What most firms don’t know: Google’s AIOs get things wrong tens of millions of times every hour. A recent New York Times analysis (subscription required) found that AIOs produce incorrect answers roughly 9 percent of the time; that sounds modest until you do the math on five trillion-plus searches a year. When your firm is on the receiving end of one of those errors, you have no way to know it happened and no mechanism to correct it.

Not appearing in the AIO at all is a problem. Getting an inaccurate one is worse. Most firms are flying blind on both.

Most firms catering to business clients have invested meaningfully in their online presence over the years by building strong websites, optimizing bios and practice pages, producing blog posts and thought leadership content, and maintaining LinkedIn profiles and directory listings. That investment was well-placed, but it was built for traditional search. AI search sizes you up differently.

Your Website Is Not the Story Anymore

When a prospect asks ChatGPT, Claude, or Gemini about your firm, the platform doesn’t experience your website the way your average water-based, carbon life form does. It pulls signals from wherever it can find them, and the sources may surprise you.

Are you sitting down? Research by McKinsey & Company found that a business’s own website represents only 5 to 10 percent of what AI search draws from when assembling an answer. The rest is third-party. To make that concrete: the same New York Times analysis found that Facebook is the second-most-cited source in Google’s AI Overviews. Reddit is fourth.

Despite doing all the right things (current directories, press releases, lawyers’ articles published outside the firm’s domain), many firms still have a thin AI profile. A firm whose information appears consistently across authoritative sources gets surfaced, but one whose information lives primarily on its own domain does not.

Google’s AIO compounds this. Those synthesized answer blocks aren’t pulling from a single source; instead, they aggregate material from across the web. If all the good stuff about your firm and its people is like Tom Hanks and Wilson on the island, the AIO answer to anyone researching you will be lean, generic, or unknown. Remember that hot prospect your pal referred? They just became some other firm’s next best client.

What AI Search Actually Rewards

Fortunately, what AI search is looking for is not a technical mystery. As a Wall Street Journal piece (subscription required) explains, it rewards firsthand experience, specificity, and content that reflects genuine, demonstrable knowledge. Not keyword stuffing and voluminous content, but actual expertise that shows up across credible sources and reads like it was written by someone (a human) who knows what they’re talking about.

Good or bad, the system can be gamed. An AI search executive at a marketing agency noted that writing a blog post that self-proclaims expertise is sometimes enough to get it surfaced in AI answers. A BBC podcast host tested the theory by publishing a post about a fictional South Dakota hot dog-eating championship and listed himself as the winner. The next day, Google’s AIO cited it as fact and ranked him among the top “hot dog-eating tech journalists.” The system treats a self-promotional blog post the way it treats a verified credential.

Firms that have been doing the right things already have the real ingredients: current directory profiles, publicized rankings and peer recognition, informative articles under attorneys’ names, and case studies and outcomes documented somewhere a search can find them. The question worth asking is whether those ingredients are visible and specific enough for the AIs to assemble into a coherent, authentic answer when someone asks.

Why AI-Generated Content Makes This Worse, Not Better

The irony here is hard to miss. The quality AI search rewards (specificity, firsthand perspective, genuine knowledge) are exactly what AI-generated content can’t produce. But the legal marketing space is flooded with vendors peddling AI content for less than what human writers charge, and plenty of firms are buying it under the assumption that more content means more visibility.

And the lawyers who fancy themselves marketers, asking AI to write their content because they think they can DIY it? Well, you get what you pay for.

Generic, templated output that reads like a description of any firm in your practice area, rather than yours specifically, is precisely what AI platforms discount. Filling your site with it doesn’t build a stronger AI footprint. It builds a weaker one because the signals AI search looks for are absent from content a model produced in seconds with no understanding of your practice. AI-generated content degrades quality signals while delivering nothing a knowledgeable human writer can’t do better.

There’s also an ethics problem that most firms wave past. Bar advertising rules vary by jurisdiction and can require claims about outcomes, practice strengths, and attorney credentials to be reviewed by someone who actually knows those rules. Courts have sanctioned attorneys repeatedly for submitting unreviewed AI output in legal filings. Marketing content may be a different arena, but the pattern is the same: fast output, minimal review, real exposure.

The Number That Puts This in Perspective

Only 16 percent of brands systematically track how they appear in AI search, according to the McKinsey report. For business-facing law firms, that number is almost certainly lower. Most have no idea what ChatGPT or Google’s AIO returns when someone searches for them, and because many AI users get their answer without clicking through to any website, those interactions leave no trace in your analytics.

So what’s a law firm to do? Start with a simple test: type your firm’s name into ChatGPT, Gemini, Claude, and Perplexity, and read what comes back. Is it correct? Is it specific to your practice and your people, or does it read like a description of any mid-size business-law firm? Does it reflect the expertise, recognitions, and outcomes your attorneys have earned, or does it return something generic and forgettable?

Referrals remain the most powerful business development tool in a relationship-driven practice. Referred prospects still vet you online before reaching out, but now AI is their trusted advisor, verifying you before they make contact.

Making sure that verification goes your way requires content written by humans who understand the law, know the target audience, and grasp the difference between sounding authoritative and being authoritative.

Coincidentally, Rain has a team of lawyer-turned-writer humans. So you know. Just sayin’.