# AI Search Is Raising Media Budgets. The Money Skips the Answer. *CMOs are winning bigger media budgets by pointing at AI search. The extra dollars go to TV, CTV and social, and reach the AI answer only through publisher pages nobody pays for.* Published: 2026-09-17 | Read time: 5 min read | Author: Smalk AI Research | Source: https://www.smalk.ai/blog/ai-search-media-budgets-money-skips-the-answer --- AI search media budgets are going up, and almost none of the increase touches the AI answer. Digiday reported on September 15, 2026 that State Farm, Mercedes-Benz and Hoka are using AI-driven consideration anxiety to argue for larger upper-funnel budgets, not for GEO. The money lands on linear TV, CTV, paid social and out-of-home. The decision it is meant to influence is made inside ChatGPT, Gemini and AI Mode, on a surface no media plan can currently buy. - The first measurable macro effect of AI search on advertising is not a new channel. It is inflation of the old ones, with the IAB lifting its 2026 U.S. growth forecast from 9.5% to 12.3% on September 10. - Brands are buying awareness upstream because there is no way to buy presence at the point of consideration inside the answer. - The awareness halo does reach AI answers, but it travels through third-party publisher pages: Muck Rack's May 2026 study puts 84% of AI citations in earned media. - Publishers therefore carry the mechanism that justifies the TV increase and capture none of the incremental budget. ## What the consideration detour in AI search is The consideration detour is the reallocation of media budget to upper-funnel channels in response to a consideration phase that has moved into AI assistants and cannot yet be purchased directly. Brands with long buying cycles, cars, insurance, running shoes, see the shift first because their customers research longest. Digiday cites a Dept survey of 2,600 U.S. shoppers in which 61% used an AI assistant for shopping research or decisions in the past three months. Ryder's 2026 e-commerce study, with 64% of 1,160 online shoppers using AI-assisted tools, points the same way. The detour is rational. When the surface that decides the sale is closed to media, the only lever left is to shape what the surface already believes about you. That is what State Farm's head of marketing means when she calls a customer pausing on the brand name inside an AI answer a win. ## Buying the harvest on Google vs seeding the answer from TV In the open-web search economy, brand was the seed and paid search was the harvest. You built awareness on TV, then captured the intent it created with a keyword bid, a click and a landing page. Every dollar of brand spend had a downstream cash register in the SERP, and share of search told you when to expect the payout. In the agentic-web economy, the harvest moved inside the answer and the cash register did not follow. A category prompt returns a synthesized recommendation built from cited pages, with no keyword auction at the moment of choice on most engines. So budgets retreat to the seed. Brands spend more on awareness precisely because they cannot spend at consideration. The old funnel had a paid layer at both ends. The new one has a paid layer at the top and a black box at the bottom. ## What Digiday reported, and what the IAB numbers actually say Digiday's September 15, 2026 report by Sam Bradley documents CMOs and agency leads at State Farm, Mercedes-Benz, Hoka, Mediaplus and Media by Mother framing AI search as a reason to fund awareness and favorability work, with Mediaplus now reporting share of model alongside upper-funnel metrics to unlock budget. One attribution needs a caveat. Digiday says the IAB credited AI-search-driven spend as part of its forecast revision; the IAB's own September 10 release attributes the 9.5% to 12.3% uplift to the Winter Olympics, the FIFA World Cup and easing economic worries, while separately reporting that 44% of buyers name adapting to AI-driven search as their top investment challenge and 76% list optimizing for AI-generated answers as their number one area of increased focus. The AI-driven share of the uplift is real but unquantified. We use the IAB's framing. ## Why brand spend really does move AI answers, and where the halo travels ### The strongest case: language models mirror brand salience The upper-funnel argument is not wishful. Gartner's Andrew Frank tells Digiday that AI makes brands more vulnerable and brand investment can shore up the cognitive gap. Google and Tracksuit's Return on Awareness study, published June 23, 2026 across 31 brands in 15 categories, found that at roughly 30% awareness each five-point gain produced about a five-point gain in share of search. Ahrefs' 2026 analysis of 75,000 brands found brand web mentions correlate with AI visibility at 0.664, three times the 0.218 for backlinks. A model trained and grounded on the web will reflect the brands the web talks about most. Awareness leaks into the answer. ### Why the halo reaches the model through publisher pages, not TV screens The leak has a path, and it is not the 30-second spot. The Google and Tracksuit data measured branded queries on Google in Australia and New Zealand between July 2024 and June 2025, not recommendations inside AI answers, and Hoka's marketing lead concedes the link from brand spend to share of model is hard to observe. What the citation data shows instead is that models read the web's editorial layer: Muck Rack's May 2026 Generative Pulse study of more than 25 million cited links across ChatGPT, Claude and Gemini found 84% of citations in earned media, and McKinsey puts a brand's own site at 5% to 10% of referenced sources. A Coco Gauff campaign influences Gemini when it becomes a review, a comparison, a ranking on a publisher's page. The TV dollar reaches the answer only by passing through the supply side, which is paid nothing for the transit. ## What the detour means for CMOs and for publishers ### For CMOs, media buyers and agencies: fund the detour, then price the destination Keep the upper-funnel increase; the salience argument holds. Then stop treating it as the whole answer to AI search. Ringfence a separate line for presence inside and beside AI answers, with citation share and mention share as its KPIs rather than borrowed TV metrics. Audit which publisher pages actually carry your brand into ChatGPT and Gemini, because that list is your real consideration inventory, and demand ad formats native to AI agents that place the brand where the recommendation is assembled instead of hoping the halo arrives. ### For publishers: the TV budget is aimed at your pages, invoice for it Every incremental awareness dollar this quarter is a bet that a publisher's comparison, review or ranking will carry the brand into an AI answer. That makes cited pages the transmission layer of the largest ad forecast revision of the year, and today they earn from it only when a human clicks, which is less and less often. Price the influence, not the visit: cited-page inventory sold to brands that need to appear where AI agents assemble recommendations is the revenue line that matches the role. Google's AI contribution pilot, reported by Digiday on September 14, 2026, pays some publishers when content significantly contributes to an answer, but with an opaque monthly figure and no brand demand behind it. It is a rebate, not a market. ## Three signals the detour is a transition, not a destination First, measurement is professionalizing: Digiday's September 15 graphic on the scramble to measure AI visibility and the IAB's 76% focus figure show buyers building the dashboards a paid layer will need. Second, demand pipes are arriving: Amazon connected its DSP to ChatGPT ads on September 10, 2026, so brand budgets now have a programmatic route toward the answer surface. Third, supply is being counted: Google's contribution pilot puts a monthly number on cited content for the first time. When measurement, demand and counted supply exist on the same surface, budgets stop detouring and start buying directly. Expect the first 2027 media plans to carry an explicit AI-answer line funded partly from this year's TV increase. ## Conclusion Hold on to this: the first thing AI search did to media budgets was inflate the channels furthest from the answer, because the surface that now decides the sale has no buyable layer and the pages that feed it have no paid one. Smalk AI builds that layer, an AI Search ad network placing native ads for AI agents so brands can buy presence where recommendations are assembled while the publishers whose pages carry the brand into the answer finally share in the budget they transmit. Watch the 2027 upfronts: the moment an AI-answer line appears next to CTV in a national media plan, AI search media budgets stop taking the detour. ## FAQ ### Why are marketers increasing media budgets because of AI search? Because consideration has moved into AI assistants where brands cannot yet buy placement, CMOs are funding awareness upstream to shape what those assistants already believe about them. Digiday reported on September 15, 2026 that State Farm, Mercedes-Benz and Hoka are all making this case internally, and the IAB lists adapting to AI-driven search as buyers' top investment challenge. ### Does TV advertising improve a brand's visibility in ChatGPT or Gemini? Indirectly. Brand salience correlates with AI visibility, but models draw mostly on third-party editorial pages: Muck Rack's May 2026 study found 84% of AI citations in earned media. TV spend reaches the answer when it becomes coverage, reviews and rankings on publisher sites, not through the spot itself. ### What is the difference between share of search and share of model? Share of search measures how often people search for a brand by name relative to its category on Google, a proven leading indicator of market share. Share of model measures how often AI assistants mention or recommend a brand in category prompts. The first has decades of validation; the second is still a dashboard metric that nobody transacts on. ### How can publishers earn from AI search when brand budgets go to TV? By selling the influence their cited pages exert on AI answers rather than waiting for clicks. Cited-page inventory offered to brands that need to appear where AI agents build recommendations converts the publisher's role in the consideration detour into revenue. Licensing rebates such as Google's contribution pilot help, but they carry no brand demand.