AI Search Decides Who to Recommend Before the Customer Knows to Ask

Traditional search marketing operates on a straightforward model: a consumer identifies a need, types a query, and the ranking system delivers options. The marketing investment goes toward making sure a business shows up when that intent is expressed. AI-powered search is rewriting the sequence, as language models don't wait for a fully formed query. They surface recommendations in the middle of a conversation, connecting a user's situation to a service the user may not have realized they needed. That shift moves the competitive battleground upstream, from ranking for declared intent to being the business the model already trusts before the question is asked. In the first four months of 2026, roughly two-thirds of US Google searches ended without a click, so the answer a user acts on is increasingly the one the machine assembles for them.

Divyarth Churiwal, Founder of Nexus Multimedia, runs a marketing and reputation management agency serving personal injury law firms, medical spas, dentists, and home service businesses. Before launching the agency in 2018, Churiwal spent two years at Qualcomm managing supply chain operations across the U.S., Europe, and Taiwan. He also operates The PR Summit, which places clients in national outlets including USA Today, Forbes, and the Wall Street Journal, an arm of the business that has become central to his AI search strategy.

"AI can recommend your business before the customer even knows they need your service," Churiwal says. He describes the shift in concrete terms through his legal clients. In traditional search, a user who's been in a car accident types "personal injury lawyer" and Google returns a ranked list. In an AI conversation, the user describes the accident, the damage, and the circumstances. The model assesses the situation, suggests seeking a personal injury attorney, and can recommend specific firms based on the circumstances and the information it finds about them. "You're creating demand from a market that doesn't even know they need your services," Churiwal says.

The recommendation starts outside your website

For Churiwal, the signal AI models weigh most heavily when assembling recommendations lives outside a business's own site, in what the rest of the internet says about it.

"Google trusts what the internet says about you, not what your website says about you," Churiwal says. "Having third-party brands and publications talking about your brand makes much more of a dent in search results and lead intake than traditional SEO." That idea has long existed in organic search: a business's reputation isn't confined to what it publishes about itself. Churiwal sees the same dynamic growing more important as AI systems assemble answers from sources across the web. AI search systems can draw on those third-party sources when putting together answers and recommendations. A business with consistent coverage across reputable publications and a strong review profile gives those systems more third-party information to work from when assembling a recommendation.

The same logic applies in reverse. A division of Churiwal's agency, DeleteGoogleReviews.com, specializes in reputation management, specifically identifying and removing unfair Google reviews that fall under the platform's own removal policies. Competitor attacks, reviews from people who were never customers, disgruntled former employees: those signals can become part of the picture an AI system draws on when evaluating a business. "AI picks up all these reviews, takes them into consideration when deciding which business to recommend," Churiwal says. In competitive verticals like med spas, where the service itself has minimal differentiation, the review profile and third-party reputation can be the deciding factor in whether a model surfaces one practice over another.

That redraws what a company's own content is for. Blog posts published on a business's own site serve a purpose for prospective clients researching the company directly, but Churiwal argues they shouldn't be expected to drive leads on their own. "The algorithm is built to give more credibility to third-party information than what's on your website," he says. "You don't want to call yourself great. You want other people to call your services great."

The shift is architectural

The website still matters, but Churiwal's team has fundamentally changed how they build them. His developers now design site infrastructure primarily for how AI models will process it, not just for how human visitors will experience it.

"What's more important is having the script on the back end," Churiwal says. "The skeleton of the website needs to be such that AI search results and language models can easily scrape that data. There's a cost involved for models to scrape data. It eats actual power and compute. So the easier it is for models to scrape the data on your website, the more likely you are to be visible to AI search systems."

For Churiwal, that makes technical accessibility part of the visibility problem. His team treats the website as infrastructure that has to be readable by people and machines at once, one part of the recommendation process: what the web says about a business and how easily the machine can understand what it finds.

Legal is the most expensive vertical in Google Ads, and in the most contested personal-injury markets a single click runs far above the category average. Churiwal estimates that top-ranked firms capture roughly 70% of the business, and everyone else fights over the remaining 30%. That changes what SEO success means. A firm can still rank for the query a client types and lose the recommendation that forms before the query is fully shaped.

The job is no longer just getting a business onto the first page for the query a customer already knows to make. It's building enough evidence across the web that an AI system can recognize the business, understand what it does, and recommend it while the customer's need is still taking shape. "The biggest mistake I see is not even doing anything about the changing landscape," Churiwal says. Traditional search asked whether a customer could find a business once they knew what to look for. AI search asks the harder question: will the system think of that business before the customer does?