“There’s been a marked shift in awareness: brands are realizing that years of hard-earned search equity are being reshaped overnight as AI moves from search engines to answer engines,” David Hayes, head of digital, Europe at the global creative collective, Forsman & Bodenfors, said in an email to Digiday.

That said, AI search optimization has become “a board-level discussion topic for many of our partners,” Hayes added.

Perhaps for good reason. Half of consumers already use AI-powered search, according to recent McKinsey research. By 2028, $750 billion of consumer spend is expected to flow through AI-powered search platforms. Marketers have seen the writing on the wall, and have started to rethink SEO strategies to be built not for humans, but for algorithms.

“It’s not the way that clients have had to – or brands have had to – structure their website in the past, and that’s kind of the distinct strategy between SEO, AEO, and GEO,” said Savannah Bishop, group director of digital platforms at Fitzco.

Reputation and credibility

Where and how shoppers are finding brands is no longer limited to Google search results. Similar to publishers, marketers have seen a shift in web traffic in light of the rise of AI search. But in a so-called zero-click landscape, marketers are grappling with how to measure visibility and thus build out an AI search strategy. 

It’s a matter of testing and learning, at least for now. 

Digiday spoke with five agency execs, and all agreed on a few methods, including prioritizing clear questions and answers via FAQs on brand sites, structuring site content around question answers rather than keywords, adding alternative text on visual assets and creating consistent brand messaging across the web — that’s everything from messaging on owned-and-operated channels to press releases and influencer campaigns. The name of the game is credibility and authority. And that’s presumably what AI systems use to determine search results, per the execs. 

“The reputation and credibility that you’re perceived as having is more and more effective than having a direct answer to a specific question,” said Chris Rigas, vp of media at performance media agency Markacy. 

In the weeds strategies

Go Fish Digital, a digital marketing agency, has started reverse engineering AI models, learning how the search platforms and AI models work, via readily available APIs. For example, one test revealed content at the top of a page is more likely to be found and used in search results, according to Dan Hinckley, chief product and AI officer and co-founder at Go Fish Digital.

Meanwhile, Forsman & Bodenfors developed AI agents trained on brand tone of voice, visibility criteria and SEO benchmarks, Hayes told Digiday. The tools help the agency score and provide feedback on client submitted articles to speed up the editorial process and maintain consistency in messaging.

Tools like Syndigo, a software platform, are helping brands create consistent product data across retailers and digital platforms for easy processing by LLMs.

From a paid perspective, Fitzco is encouraging clients to invest in Google’s Performance Max, as the Google Ads campaign is eligible for AI Overviews, said Bishop. So even if a user doesn’t click on a search result, there’s a chance the ad will be shown within an AI-generated summary. 

“What we’re finding is that language models are acting very much like people do. They just want the information quickly and efficiently,” said Hinckley.

New measurement and attribution woes

Marketers are testing, but there’s a black box nature to the LLMs making it difficult to determine how an AI chatbot makes a recommendation.

Agencies are measuring things like how often a client’s content is surfaced in zero-click environments, investing more in digital marketing and SEO platform tools — or more plainly, looking at web traffic before and after AI search strategy changes.

“It’s pretty tough to get a grasp on what is and isn’t moving the needle, because the analytics are not always there,” said Markacy’s Rigas. 

Black-hat tricks

While black-hat tricks and SEO hacks may sound tempting, agency execs say it’s best to play the long game. Reddit has been a significant source for AI answers — perhaps so much so that last month, Reddit moved to sue Perplexity for scraping data to train its AI system. 

That said, agency execs like Rigas, said there’s been rumblings of brands directly writing posts to promote products to get picked up by LLM crawlers. Go Fish Digital’s Hinckley pointed to an early SEO “cloaking” trick in which marketers add text that is invisible to users but visible to the language model to influence its output.

In both cases, agency execs caution against quick fixes, noting that AI-powered chatbots may eventually be taught to recognize and address such tactics.

“We’re firmly in a test-and-learn phase, and it’s important that the industry approaches this moment with caution,” said Forsman & Bodenfors’ Hayes. “There’s no shortcut or ‘hack’ to AI visibility and attempts to game the system will likely backfire, just as they did in the early days of SEO.”