How ChatGPT and Gemini Monetization Will Reshape Paid Media Services in 2026

Written & peer reviewed by 4 Darkroom team members

TL;DR: ChatGPT and Gemini are rolling out paid ad products, and most paid media services are not structured to buy, optimize, or measure them. The problem is not that a new channel exists. The problem is that conversational AI ads collapse the boundary between organic visibility and paid placement, and most teams treat these as separate functions run by separate people with separate budgets. Brands that build a unified paid media strategy for AI placements now will capture favorable CPMs before auction density rises. Those that wait will pay a premium to enter a market their competitors already understand. At Darkroom, we build paid media programs that integrate new ad surfaces into existing portfolio allocation models rather than bolting on another siloed channel.

The Ad Products Nobody's Buying Yet

The media buyer pulls up the weekly performance deck. Google Ads, Meta, TikTok, maybe Amazon DSP. Four platforms, four attribution windows, four dashboards that disagree with each other. The meeting takes 45 minutes. Nobody mentions ChatGPT. Nobody mentions Gemini. And that silence is where the next wave of paid media market share is being decided.

OpenAI confirmed its advertising plans for ChatGPT in early 2025 and began testing sponsored responses with a small group of advertisers. Google has already rolled AI Overview ads into Search and Shopping campaigns in the US, placing sponsored results directly inside Gemini-generated answer panels. Perplexity, the AI search engine processing over 15 million queries daily, launched sponsored answers in late 2024 with brands like Nike, Marriott, and Universal. These are not beta experiments. They are live ad products with real inventory, real pricing, and almost no competition for placement.

The volume of commercial queries flowing through AI interfaces is growing at a rate that makes early search engine adoption look slow. But the advertising infrastructure around these queries is 18-24 months behind the user behavior shift. That gap is where the opportunity sits for brands with paid media management teams willing to run the experiments nobody else is running.

What Conversational AI Ads Actually Look Like

Conversational AI ads are not display banners inside a chat window. They are sponsored content woven into the AI's response itself. That distinction matters because it changes every assumption paid media teams carry about creative formats, auction mechanics, and measurement.

Google AI Overview Ads place Shopping and Search ads directly inside the AI-generated summary panel that appears above organic results. When a user searches "best running shoes for flat feet," Gemini generates a synthesized answer, and sponsored product cards appear within that answer alongside organic citations. The format inherits existing Google Ads campaigns, which means brands already running structured Google Ads architecture have a head start. But the optimization logic is different. Click-through rates in AI Overview placements run 2-3x higher than standard Shopping ads because the recommendation feels earned, not placed.

ChatGPT Sponsored Responses work differently. OpenAI's approach embeds product recommendations and branded information cards directly within ChatGPT's conversational output. A user asking "What CRM should a 50-person SaaS company use?" might see a sponsored recommendation for HubSpot integrated into ChatGPT's answer, clearly labeled but formatted to match the surrounding response. There is no list of 10 results. There is one answer with a sponsored component. That is a fundamentally different competitive dynamic than keyword auctions.

Perplexity Sponsored Answers appear as branded follow-up questions in the side panel and as cited sources within the AI's response. A user researching hotels in Tokyo might see "Sponsored by Marriott: How do Marriott Bonvoy points work for Tokyo stays?" as a suggested follow-up question. According to Reuters reporting on Perplexity's ad system, the platform uses brand-provided targeting data to match sponsored content with relevant queries, creating a hybrid between search ads and native content.

The common thread across all three: the ad is inside the answer. Not next to it. Not above it. Inside it. That collapses the distance between organic visibility and paid placement in a way that traditional search never did. And it means your paid media channel allocation strategy needs a new column.

Why This Is Not Just Another Channel Launch

Every 18 months, a new ad platform gets the "this changes everything" treatment from marketing media. Most of the time, it doesn't. Pinterest ads, Snapchat ads, Reddit ads: useful for niche applications, but none of them restructured how brands think about paid media strategy. AI ad placements are structurally different, and the difference comes down to one thing: the AI response is replacing the search results page for an increasing share of commercial queries.

Google processes roughly 8.5 billion searches per day. As of early 2026, AI Overviews appear on over 40% of those queries, according to Search Engine Land's analysis of Google's rollout data. That means billions of daily queries where the first thing a user sees is an AI-generated answer, not a list of blue links. For many of those queries, the user never scrolls past the AI Overview. The click-through to a website happens inside the AI response or it doesn't happen at all.

ChatGPT surpassed 400 million weekly active users in early 2025. A growing share of those sessions involve product research, brand comparison, and purchase consideration queries that would have been Google searches two years ago. When the user asks ChatGPT "What's the best protein powder for a 40-year-old who lifts 4 days a week?" instead of searching Google for "best protein powder," the entire paid search funnel for that query disappears. Unless the brand has a presence inside ChatGPT's response.

This is not a channel addition. It is a channel migration. The queries are the same. The user intent is the same. The venue changed. And paid media services that only know how to buy Google keywords and Meta audiences are structurally unprepared for where that intent is moving. Understanding how content feeds power AI search engines is the first step toward seeing the shift clearly.

The Early Mover CPM Advantage

The economics of new ad platforms follow the same curve every time. Early inventory is abundant and underpriced because advertiser demand has not caught up with user adoption. Then holding companies shift budget allocation. Then mid-market brands follow. Then CPMs normalize or spike. The window between "live" and "expensive" typically lasts 12-18 months.

Facebook Ads in 2012-2013 had CPMs under $3 for newsfeed placements that cost $12-$18 by 2016. TikTok Ads in 2020-2021 had CPMs of $4-$6 for in-feed placements that now average $8-$14 in competitive verticals. The brands that built their acquisition engines during those windows, and built the internal knowledge to optimize on those platforms, had compounding advantages for years after.

AI ad platforms are in this exact window right now. Perplexity's sponsored answers are available to a limited set of advertisers. Google's AI Overview ads inherit existing campaign structures but very few brands are actively optimizing for them. ChatGPT's ad products are in early testing. The CPM floor has not yet been set by competitive auction pressure.

But there's a second advantage that goes beyond pricing. Brands that start buying AI ad placements now accumulate proprietary performance data that late entrants won't have. Which query categories convert? What does a "good" conversational ad look like? How do AI placements interact with existing search and social spend? These are questions that can only be answered with live campaign data, and that data compounds over time. The shift from backward-looking attribution to predictive measurement depends on having enough signal to model future performance, and early AI placement data feeds that model.

AEO: Where Organic and Paid Converge

Answer Engine Optimization is the practice of structuring content so that AI systems cite it when generating responses. It combines structured data, definitional clarity, entity associations, and recency signals. Until recently, it was a purely organic discipline. A content strategy play. That is changing fast because paid placements inside AI responses sit right next to organic citations, and the boundary between the two is thinner than anything we've seen in search.

When Google shows an AI Overview with a sponsored product card and three organic citations, the user sees one blended answer. The sponsored card and the organic citations occupy the same cognitive space. A brand that appears in both the paid placement and the organic citation has dramatically higher click-through and trust signal than a brand appearing in only one. This is why paid media strategy and content strategy can no longer live in separate departments reporting to separate VPs with separate KPIs.

Darkroom's paid media agency approach builds AEO into the media plan rather than treating it as a separate organic initiative. The operational implication is straightforward: the same team optimizing Google Ads campaign structure needs to know what structured data is on the landing page, what schema markup is feeding AI crawlers, and whether the brand's content shows up in AI-generated answers for the queries they're buying. A paid media agency playbook that doesn't include AEO is a playbook with a growing blind spot.

The convergence also means measurement gets harder. When a customer sees your brand cited organically in a Gemini response, then clicks a sponsored Shopping ad in the same response, which channel gets credit? When ChatGPT recommends your product in an unpaid citation and the user later searches your brand name on Google, does the AI citation get attributed? Traditional measurement frameworks break here. The teams already investing in geo experimentation for paid media measurement have the infrastructure to isolate these effects. Teams relying on platform-reported attribution don't.

What Changes Operationally for Paid Media Teams

Creative formats need a complete rethink. Search ad copy is built for 30-character headlines and 90-character descriptions competing against nine other results. Conversational AI ad copy exists inside a paragraph written by an AI. The sponsored content needs to feel like a natural extension of the answer, not an interruption. This is closer to native advertising than search advertising, and most performance creative agency teams don't yet have templates for conversational ad formats. The brands that develop those templates first will set the creative standard for the category.

Keyword targeting becomes query-intent mapping. In traditional search, you bid on keywords. In conversational AI, the user asks a question, and the AI decides whether your ad is relevant to the answer it is generating. The targeting mechanism is closer to contextual matching than keyword bidding. This means the quality of your product feed, your structured data, and your brand's entity associations in knowledge graphs matter more than your keyword list. A brand with strong Amazon marketing product data infrastructure (clean titles, structured attributes, complete backend keywords) will find that the same data quality translates directly to AI ad relevance.

Attribution gets messier before it gets cleaner. AI ad placements don't have the same click-path tracking that search ads do. A user might see a sponsored recommendation in ChatGPT, not click it immediately, then search the brand name on Google an hour later. The AI placement influenced the purchase but won't show up in any last-click model. Teams that rely on platform-level metrics alone will undervalue AI placements and underinvest in them. Teams using incrementality testing and multi-layer attribution stacks will be able to isolate the true lift.

Budget allocation needs a test-and-learn line item. The right approach is not to move 20% of Google spend into ChatGPT overnight. It is to carve out 5-10% of paid search budget as a dedicated AI placement testing budget with its own measurement framework. Run it for 90 days. Measure brand search lift, assisted conversion rate, and cost per qualified visit. Then scale based on incremental contribution, not platform-reported ROAS. Darkroom builds ROAS frameworks that account for new channels where platform reporting has not yet matured.

The PPC Optimization Playbook for AI Placements

Running paid media inside AI interfaces requires a different optimization playbook than standard PPC campaign optimization. Here is the framework Darkroom recommends for brands entering the space.

Phase 1 (Days 1-30): Infrastructure and data foundation. Audit your product feed quality, structured data markup, and entity associations. Run a baseline measurement of brand search volume and direct traffic before activating AI placements. Set up a holdout geography for incrementality testing. If your Google Ads campaign structure is not segmented by intent type (brand, non-brand, competitor), fix that first, because AI Overview ads inherit your existing campaign architecture and poor structure compounds in the new format.

Phase 2 (Days 31-60): Controlled testing. Activate Google AI Overview ad eligibility on your highest-performing Shopping and Search campaigns. If you have access to Perplexity's sponsored answers program, start with your top 5 product categories. Begin tracking AI-specific KPIs alongside traditional metrics: AI impression share, conversational ad engagement rate, and post-AI-exposure brand search lift. Compare these against your existing Google versus Meta budget allocation to understand where AI placements fit in the portfolio.

Phase 3 (Days 61-90): Measurement and scaling decisions. Run the incrementality analysis. Compare brand search volume in test markets versus holdout markets. Calculate cost per incremental conversion from AI placements. If the incremental CPA is at or below your blended target, scale. If it is above target but within 20%, optimize creative and targeting before scaling. If it is more than 20% above target, pause and investigate whether the measurement window is too short for the purchase cycle.

The Structural Gaps in Most Paid Media Services

Most paid media services are structured around a specific set of platforms: Google, Meta, TikTok, Amazon. The team has buyers who specialize in each platform's ad manager, creative teams that produce assets for each format, and reporting dashboards that show performance by channel. That structure works for the current ad ecosystem. It does not work for what is coming.

The first gap is organizational. AI ad placements require collaboration between paid media, SEO, content strategy, and data engineering. The paid buyer needs to understand structured data. The content strategist needs to understand bid mechanics. The data engineer needs to build attribution models that account for AI-influenced conversions. In most agency structures, these functions report to different people, operate on different timelines, and have never worked on the same campaign. That isolation is the gap.

The second gap is measurement. Most paid media agency playbooks measure new channels by importing them into existing reporting frameworks. But AI placements don't fit neatly into platform-reported ROAS models. They produce their value through brand lift, consideration-phase influence, and query-level visibility that doesn't map to a click-through conversion. If the measurement framework can't capture these effects, the channel will always look like it's underperforming, and budget will flow to channels that report better but may contribute less.

The third gap is speed. The brands that captured the Facebook Ads window in 2012 and the TikTok Ads window in 2020 had one thing in common: they made allocation decisions before the data was complete. They accepted 60% confidence and iterated. The brands that waited for "proven ROI" entered markets where CPMs had already tripled and the learning curve advantage had been captured by competitors. AI ad placements are in the same 60% confidence window right now. Paid media services structured around proven playbooks will miss it. Services structured around portfolio-level experimentation won't.

How to Evaluate Your Paid Media Agency's AI Readiness

If you work with a paid media management partner, ask these five questions. The answers will tell you whether they're prepared for where digital advertising trends are heading or whether they're optimizing for a media landscape that is already shifting underneath them.

The difference between the two columns is the difference between a reactive agency and one that operates at the frontier. According to eMarketer's 2026 digital advertising forecast, AI-powered ad formats will capture 8-12% of total search ad spend within 18 months. That is not a rounding error. It is a structural reallocation that will reward agencies and brands already positioned to capture it.

Darkroom's paid media team builds AI placement testing into the portfolio allocation model from day one. The same measurement infrastructure that powers channel allocation decisions extends to AI placements, ensuring they are evaluated by incremental contribution rather than platform-reported metrics that don't yet exist at maturity.

Frequently Asked Questions

What are conversational AI ads?

Conversational AI ads are sponsored placements that appear within AI chat interfaces like ChatGPT, Gemini, and Perplexity. Unlike traditional search ads that display alongside a list of blue links, conversational AI ads surface as sponsored product recommendations, inline citations, or branded answer cards within the AI's natural-language response. The format is closer to a recommendation from a trusted advisor than a banner ad on a results page.

How do ChatGPT ads differ from Google Search ads?

ChatGPT ads appear inside a conversational response rather than alongside a list of search results. The user asks a question, receives a synthesized answer, and sees a sponsored recommendation embedded within that answer or appended below it. There is no equivalent of position 1 through 10. There is one answer, and the ad either appears inside it or it doesn't. This makes the format closer to a featured placement than a keyword bid auction.

What is Answer Engine Optimization or AEO?

Answer Engine Optimization is the practice of structuring content so that AI systems like ChatGPT, Gemini, and Perplexity cite it when generating responses. AEO combines structured data, definitional clarity, entity associations, and recency signals to increase the probability that an AI model references your brand or content. It sits at the intersection of organic SEO and paid media because being cited organically in an AI response and paying for a sponsored placement within that response are converging into the same visibility strategy.

Are AI ad placements available to buy right now?

Google's AI Overview ads within Gemini-powered search are already live for Search and Shopping campaigns in the US. OpenAI has announced ad plans for ChatGPT and begun testing sponsored responses with select partners. Perplexity launched its sponsored answers program in late 2024 with brands like Nike and Marriott. Availability is expanding rapidly, but most formats are still in limited beta or early rollout. Brands that establish relationships with these platforms now will have data and optimization advantages when the formats open to broad auction.

Why are early CPMs lower for AI ad formats?

Early CPMs are lower because advertiser demand has not caught up with available inventory. AI platforms are growing their user bases faster than their ad sales teams can fill the supply. This mirrors the early days of Facebook Ads in 2012-2013 and TikTok Ads in 2020-2021, when brands that entered early locked in CPMs 40-60% below what the same inventory cost 18 months later. Once major holding companies shift budget allocation toward AI placements, auction density will increase and CPMs will normalize.

How should brands allocate budget for AI ad placements?

Start with 5-10% of your existing paid search budget allocated to AI placement testing. Run it as a dedicated experiment with its own measurement framework, not as an extension of your Google Ads account. Track cost per qualified visit, assisted conversion rate, and brand search lift as primary KPIs. Treat the first 90 days as a data-collection phase rather than a performance phase. The goal is to build a proprietary dataset on what conversational ad formats convert for your category before competitors have the same data.

Will AI ads replace Google Search ads?

AI ads will not replace Google Search ads. They will layer on top of them and absorb a growing share of the discovery-phase queries that currently drive top-of-funnel search spend. High-intent transactional queries will stay in traditional search for the foreseeable future. But informational and commercial investigation queries are already migrating to AI interfaces. Brands that only buy keywords on Google will lose visibility on the queries that shape consideration before a purchase decision is made.

What paid media services are needed for AI advertising?

AI advertising requires a combination of paid media buying, content strategy for AEO-optimized assets, structured data implementation, conversational ad creative production, and incrementality measurement to isolate AI placement lift. Most paid media agencies lack the content and structured data capabilities, while most SEO agencies lack the media buying infrastructure. The gap creates an advantage for full-service teams that operate across both disciplines within a single portfolio allocation framework.

Looking for a paid media strategy that includes AI ad placements before your competitors lock in the early CPMs? Book a call with Darkroom to build an AI placement testing framework into your existing media portfolio.