By the numbers: How marketers are building the infrastructure for AI search

In some sense, they don’t have a choice. Nearly one in five consumers have already replaced traditional research with AI, according to a recent SponsorCX survey. Call it a fear of missing out (FOMO). Marketers are spending on tools promising AI visibility. They’re optimizing web pages for AI bot crawlers. Marketing budgets now account for GEO efforts. All the while, there’s little known about the LLMs’ decision making processes.

It’s left marketers with a fragmented infrastructure and the numbers prove as much.

AI visibility on the P&L

While AI visibility could largely be considered an organic play, there are still dollars required to fund things like content production and earned media. Marketers are spending, but not all at the same rate. They’re routing about 24% of search or content budgets to AI visibility, according to a recent survey from Fractl, a content marketing and digital PR agency. The survey added that 82% of marketers have allocated at least some budget, and 43% are spending more than 20% on AI visibility. Meanwhile, 18% have allocated nothing at all. 

When it comes to who’s spending, Fractl reports that SEO (31%) and performance marketing (32%) teams shell out the highest share of their budgets on AI search visibility. Content (20%) and brand (21%) spend the least.

A battle for eyeballs

Increasingly, shoppers are looking to AI chatbots for initial research. They don’t turn to search until closer to purchase, per Similarweb’s AI visibility study.

Of survey respondents, 35% reported using AI search to discover products or get initial ideas. In comparison, 13.6% started with traditional search engines. When it gets closer to purchase, the numbers are neck and neck. 24.3% of those surveyed said they use AI search while 22.1% choose traditional search. LLMs are becoming another brand discovery layer, but traditional search engines still get the clicks. 

The struggle to scale content

With LLMs becoming more of an awareness channel, more marketers are building out content strategies around it. 10Fold is a B2B tech marketing and PR agency. Its recent AI search study states 52% of B2B marketers cite AI-generated search and answer engines as the most effective distribution channel. That’s in comparison to 29% of those who report organic search as the top content distribution channel. 

And research from Scrunch, an agent experience platform, and Scribewise, a B2B marketing agency, makes a similar point. Per a recent AI search survey, 42% of marketers are scaling content production and 73% have invested in the tools to monitor AI visibility. 

Disconnect in AI infrastructure

For all the infrastructure marketers are building around AI visibility, they’re still working out the kinks. 

There’s a gap between strategy and execution. The numbers from Scrunch, Scribewise and 10Fold prove that as well. The majority of 10Fold’s respondents (41%) said only 25% to 49% of their content has been created or optimized for AI discovery. 

Scribewise’s figures take it a step further, noting 67% of respondents say their brands or clients appear less frequently in AI-generated answers than they’d like. Meanwhile, 45% of respondents say they’re testing their brand’s visibility across AI platforms. 

That confidence, however, starts to break down with some digging. About 71% aren’t tracking share of voice against competitors. Another 70% aren’t monitoring sentiment toward their brand and 67% aren’t analyzing AI bot traffic. 

Platform fragmentation 

Marketers aren’t to blame for the confusion. The decision making process behind LLMs is largely a black box. Strategies that work to get a brand cited or mentioned in ChatGPT may not work for Google’s LLMs. For marketers that have geared their entire SEO strategies toward Google search, the current slate of LLMs requires a new, more diverse playbook. 

Case in point: OpenAI’s ChatGPT pulls from an average of 15.4 sources with Wikipedia and Reddit as predominant buckets, according to Semrush’s AI visibility index. Google’s AI Mode pulls from about 11.4 sources, leading with places like Facebook, Instagram and Yelp. YouTube is the single biggest source for Google AI Overviews. Meanwhile, Gemini averages 3.3 sources, favoring entertainment and commerce sites like YouTube, Wikipedia and e-commerce leads, per Semrush.

That doesn’t account for the AI citations that stem from news organizations. At least 27% of AI citations are drawn from journalism, per insights from 5W comms firm. When it comes to time-sensitive queries, that figure rises to 49%.