“AI Overviews” began topping Google search results in May 2024. After more than a year, it’s become clear that AI has fundamentally changed the role of search in the strategies of brand and reputational managers. The upshot: The AI revolution in search is making earned media, branded content, and social media marketing more important than ever.

We’re all now familiar with Google’s AI output. Search has moved on from just a listing of links deemed relevant by the SERP algorithm to an AI “summary” that sits at the top of the page.

In effect, each query about a brand – or including a term or concept relevant to a brand and its category – now generates an instantly compiled, Wikipedia-like narrative that the search user sees first. Adjacent to this narrative are a handful of links to the content used as reference material. This narrative is constructed according to Gemini AI’s (or some other AI engine) understanding of the query (aka “prompt”) and is compiled from sources that the AI engine deems credible and authoritative.

As a result, communications and marketing professionals have moved from a world focused on page rankings to one driven by the Large Language Models (LLMs) used by AI. A recent blog post by VC tech powerhouse Andreesen Horowitz noted:

“In the SEO era, visibility meant ranking high on a results page. Page ranks were determined by indexing sites based on keyword matching, content depth, and breadth, backlinks, user experience engagement, and more. Today, with LLMs like GPT-4o, Gemini, and Claude acting as the interface for how people find information, visibility means showing up directly in the answer itself, rather than ranking high on the results page.”

In this new paradigm, it’s no longer about first-page visibility or click-throughs.  It’s about reference rates – how often your brand or content is cited or used as a source in AI-generated summaries. That said, achieving constructive brand references in AI summaries is subject to multiple vagaries.

First, the wording of the prompt determines how AI treats a query. Though the prompt may include a core term or concept highly relevant to a category of brands, there’s no guarantee that any specific brand will show up in the summary. That depends on how AI categorizes and “thinks” about the prompt.

Even if the source material cited alongside the AI summary mentions specific brands, it’s unlikely to be seen by most users. A Pew Research report found that Google users who encountered an AI summary clicked on a traditional search result link in only 8% of all visits. Those who did not encounter an AI summary clicked on a search result nearly twice as often (15% of visits). That translates into a reduced opportunity for search users to “discover” brands or branded content in the source material of AI summaries.

Adding to the vagaries of AI summaries are the inherent differences in LLM models being used by AI services. Beyond Google, it’s relatively easy for users of the Safari browser, for example, to access ChatGPT, Claude, Perplexity, and other LLM search engines. All these models are continuously evolving, not only reflecting changes by the AI firms but also their models’ “learning” from prior searches. Keep in mind that LLMs are in a sense organic, always acquiring new information and context as they process prompts.

It’s also worth noting that major news sites are tightening AI engines’ access to their content. Dozens of news sites have reached content licensing deals with the top AI players (News Corp., Financial Times, Associated Press, to name a few).  But news sites are locking out AI engines that don’t have deals, which can limit the range of content used as references in an AI summary and displayed as links.

The Wall Street Journal reports that publishers are “stepping up efforts to protect their websites from tech companies that hoover up content for new AI tools.” The clampdown comes as “search traffic has dropped precipitously for many publishers, who are bracing for further hits after Google began rolling out AIMode, which responds to user queries with far fewer links than a traditional search.”

An important trend to watch is how much access social media sites end up giving to external AI services for training and citation. X and Meta are said to be using user-generated content to train their LLM models, named Grok and Llama respectively.  Reddit struck a deal with Google and some other AI companies for selected access but recently complained that some AI players are breaching the platform’s policies by scraping archived Reddit information.

For brand and reputation managers, the AI search business model looks far different from that of traditional search. Search engines such as Google have monetized user traffic through ad sales. But AI LLMs are paywall-subscription services mostly. As such, they prioritize the utility and quality of their output for the user.

Ads are still displayed on the first page, but there’s little incentive for AI to feature specific third-party content in the accompanying summary. Pay-for-reference could raise questions about the credibility and reliability of the output. A model that would monetize AI search results may yet appear, but, as Andreesen Horowitz notes, “the rules, incentives, and participants would likely look very different than traditional search.”

How should marketers and reputations think about AI search within their overall communications strategies? In many ways, the new era of AI search argues for a back-to-basics approach:

1. Treat AI summary output as another form of media, one based on derivative content. Think of LLMs as combining all the editorial and reportorial functions needed to generate a media article based on the prompt. Once again, the potential “article” will constantly evolve depending on the LLM’s learning and the access it has to searchable information. Treating LLM output as another form of media means continuously monitoring and analyzing AI summaries across a range of test prompts using brand words, key terms, and topics of high relevance to determine how LLM’s are portraying your company, your competitors, and your industry.
2. Pay close attention to AI summary reference sources. In monitoring LLM output, note which sites are being used as sources for the search summary. Which news media, industry trade publications, newsletters, or corporate site material are factors? In some cases, certain industry trade publications may be having a disproportionate impact on AI output and are thus worth targeting for earned media or thought leadership content.
3. Step up activities for driving brand visibility and placing branded content. AI search hasn’t changed this: Clear, consistent, and effective brand visibility across channels remains important – more so, in fact. Core activities around earned media, corporate news, social media redistribution and paid campaigns, online video and other forms of thought leadership help ensure that LLMs have a critical mass of your content to learn from and draw on.
4. Structure your owned content to make it more digestible by LLMs. LLMs don’t “rank” content in the same way as Google, crawling and indexing web pages.  Instead, they generate the narrative based on pre-trained data, considering word frequency, contextual relevance, and similar content. As a result, “generative engines prioritize content that is well organized, easy to parse, and dense with meaning – not just keywords,” notes Andreesen Horowitz. “Phrases like ‘in summary’ or bullet-point formatting help LLMs extract and reproduce content effectively.”

Bottom line: AI tech is revolutionizing search and impacting the media as well as the online experience more broadly, but tried-and-true methods of brand building and reputation management can help companies capitalize on these changes.