How will AI agents change online advertising?

AI agents take over the research step of shopping: comparing products, reading reviews and checking prices. That is where search ads, review sites, affiliate links and open-web display ads earn most of their money, so those are the most exposed. Advertising that shapes preferences, such as creator content and video feeds, and advertising at the point of sale, such as sponsored listings inside a store, look likely to hold up or gain. Pricing is also likely to move from paying per click towards paying per sale.

Is Google search advertising declining?

Not yet in total. Alphabet's Search revenue grew 12% year on year in Q3 2026 to $63.35 billion. But it was flat from the previous quarter while search queries grew at their fastest pace in five years, which means each search is earning less. Management said lower monetisation of AI Mode, where queries run about three times longer, weighed on growth. Paid clicks rose 8% and cost per click 4%. Ads that Google places on other websites fell 14% in impressions.

Which companies benefit from AI agents in advertising?

In our AI-assisted view, the strongest position belongs to companies that both shape what people want and own the sale. Meta grew ad revenue 26% in Q3 2026 with price per ad up 14%, and is adding commission fees through its Muse agent. Retail media owners such as Amazon and Walmart are also well placed. Platforms that influence people but send them elsewhere to buy, such as Pinterest, are being pushed towards cheaper performance pricing. The names are worked examples only.

What is happening to publishers and Reddit's traffic from Google?

Chartbeat data shows Google Search referrals to publishers fell about 40% in the year to July 2026, while AI chatbots sent just 0.01% of publisher traffic. Reddit's US logged-out daily users fell 4% in Q3 2026, and it warned total daily users could fall by a low double-digit percentage in Q4 as Google's AI answers send fewer visitors. Reddit's revenue still grew 53%, because most of its money comes from logged-in users.

What does owning the sale mean in advertising?

A platform owns the sale when the purchase happens on its own service, so it can see which ads led to which sales and charge for results. Amazon, Walmart and TikTok Shop own the sale through their stores and checkout. Meta is building towards it with conversion measurement and commission fees on purchases made through Muse. Platforms that send shoppers to another website to buy have less proof of results and less pricing power.

← Writing

Theme Deep-Dives

How AI Agents Change Online Advertising, and Which Stocks Win

3 November 2026 · 16 min read

AI agents are moving advertising money away from the research step of shopping and towards two ends: where preferences are formed, and where the sale happens. Search ads, review sites, affiliate links and ads on the open web are the most exposed, because agents now do the comparing. Creator content, video feeds and ads inside stores look likely to hold up or gain. The Q3 2026 results already show the shift, and the companies best placed are those that both shape what people want and own the purchase.

This is the third post in a series. The first covered the October 2026 AI stocks selloff, and the second covered who is likely to win consumer AI adoption. This one looks at what agents mean for the roughly $1.3 trillion advertising market. The framework, scenarios and stock groupings are our AI-assisted view. They are neither a forecast nor advice.

The ad market, and the part agents take over

The overall market is still growing. WPP Media expects global ad revenue to rise about 9% in 2026, to around $1.3 trillion, and WARC puts growth at nearly 12%. Social media, search and retail media make up about two-thirds of it, and retail media alone is about 15%.

What changes is the mix. Advertising has always paid to influence a decision. When an agent makes more of the decision, the stages of the shopping journey are affected differently:

Stage	Today	With agents	Likely direction
Preference (video, social feeds, creators, TV)	Ads and creators shape what people like	People still watch and scroll, and they tell their agent what they prefer. Being the brand people ask for matters more when the agent narrows the choice	Holds up or gains
Research (search, reviews, comparison sites, affiliate links, open-web display)	Brands pay to appear while people compare	The agent does the comparing, so fewer people see search results, review pages or affiliate links	Most exposed
Purchase (retail media, sponsored listings)	Brands pay to appear where people buy	Brands pay to be the agent's pick, through sponsored prompts, commission fees and ads inside the store's own agent	Gains, priced more per sale

What the Q3 2026 results show

Platform	Latest ad growth	What it owns	What the numbers say
Walmart Connect (US)	+43% (quarter to July, excluding Vizio)	The store, plus Vizio TV	Retail media growing fastest, at a retailer that has opened its site to agents
Meta	+26%, price per ad +14%	Preference, and building towards the sale	Advertisers paying more for feeds and creators
Amazon	+12% in Q3, about +18% across Q2 and Q3; high teens guided for Q4	The store	Prime Day moved into Q2, so the two quarters together give the clearer picture
Google Search	+12% (Q2: +16.8%)	The research step	Queries at a five-year high, but revenue flat from Q2
Pinterest	+12%, price per ad −4%	Preference only	Strong engagement, weaker pricing
YouTube	+9%	Preference, with checkout being added	Record share of US TV viewing, but slowing ad growth
Google Network (ads on other sites)	+1%, impressions −14%	Neither	The open web is shrinking
Reddit	+54% (small base)	Preference among logged-in users	Revenue rising while US visitors from Google fall

The results point to five findings.

1. Google keeps the user but earns less per search

Alphabet's Search revenue grew 12% to $63.35 billion, but it was flat from the June quarter, when a year earlier the same step added 4.4%. Over the same period, search queries grew at their fastest pace in five years. Paid clicks rose 8% and cost per click 4%. Management said lower monetisation of AI Mode, where queries run about three times longer than classic searches, weighed on growth, and the CFO expects a few more points of pressure in Q4.

Google is responding by putting more ads inside AI Overviews and AI Mode, including ads that run across several turns of a conversation in the US. It says sponsored and organic product listings in AI Mode already get similar click rates.

The bigger signal is in how Alphabet is paying for its buildout. Free cash flow was negative $9.1 billion in Q3, the second negative quarter in a row, against capex guidance of $195–205 billion for 2026. In June, Alphabet raised about $90 billion in equity, including a $10 billion private placement with Berkshire Hathaway, and said the public offerings would fund "capital expenditures to scale AI infrastructure and global compute." In Q3 it also issued $28.7 billion of bonds. Search is no longer paying for the AI buildout on its own.

2. Websites lose the visitor

Chartbeat's data shows Google Search referrals to publishers fell about 40% in the year to July 2026, after a 22% fall the year before. AI chatbots sent just 0.01% of publisher traffic. Smaller publishers lost about 60% over two years.

Reddit shows what this looks like in a company's results. Its US logged-out daily users fell 4%, search referrals were falling by more than a million a week in October, and it warned total daily users could fall by a low double-digit percentage in Q4. AI answers draw heavily on Reddit's content, but they send fewer people to it. Revenue still rose 53%, because US revenue per user rose 53% and most of the money comes from logged-in users. Reddit is now testing a more logged-in experience, partly to make scraping harder. Its "other" revenue line, which may include data licensing, was $48 million, about 5% of the total: licensing has not replaced traffic.

3. Influence is gaining

Meta's ad revenue grew 26%, with impressions up 10% and price per ad up 14%. Instagram time spent per user hit an all-time high. The fastest-growing product is partnership ads, where brands pay to promote creators' posts, which grew 97% to a $15 billion annual run rate. Across the US market, IAB estimates creator ad spend rising from $37.1 billion to $43.9 billion in 2026, with paid promotion of creator content up 48%.

Pinterest shows how far discovery has moved ahead of search. About 70% of search journeys on Pinterest begin with a recommendation made before the user types anything, and about 80% of searches in its AI assistant are about buying. It said popular consumer agents from other companies have had no visible effect on its North American engagement.

4. But influence pays best when the platform owns the sale

Pinterest's engagement is strong, but its ad pricing fell 4%. Its growth came mainly from performance advertisers paying per click, and it sends shoppers to retailers' own sites to buy. Part of the weakness is an EU rule change that hit cross-border sellers and took about 3 points off growth, but the contrast with Meta is clear.

YouTube makes the same point. It reached a record 15.8% of US TV viewing time in August, yet its ad growth slowed to 9%. Its response is to earn from influence through commissions: creator commission payouts doubled, its shopping affiliate programme is expanding from 15 to 35 markets, and it has added two-click checkout on TV screens.

TikTok Shop, which owns the checkout, is growing fastest of all. Momentum Works estimates its global sales at $50.3 billion in the first half of 2026, up 92%, with about 40% coming from videos and 8% from live streams. In October TikTok added an AI shopping assistant and Buy Direct, a native checkout inside the app.

5. Stores are turning their own agents into ad space

Amazon's ad revenue was $19.6 billion in Q3. Growth slowed to 12% because Prime Day moved into Q2, but across the two quarters it grew about 18%, faster than Google Search, and Amazon expects high-teens growth in Q4. Amazon now places sponsored "prompts" inside Alexa for Shopping, and existing sponsored ad campaigns are automatically eligible to run there. Walmart Connect grew 43% in the US, excluding Vizio, and Walmart's global advertising grew 38%.

Outside the stores, ChatGPT ads run in 89 countries with more than 100,000 advertisers, though independent estimates put their click rates at under 1%, against several percent for search ads. Ads inside an answer get fewer clicks, which is another reason pricing is likely to move towards paying for the sale.

The framework: influence and the sale

Putting the findings together gives a simple map.

	Owns the sale	Does not own the sale
Strong influence	Strongest position: Meta (building towards it with Muse), TikTok Shop, Amazon's video and Prime Video ads	Pushed towards cheaper performance pricing: Pinterest, YouTube's ad business
Weak influence	Retail media and checkout: Amazon's sponsored listings, Walmart, Shopify	Most exposed: open-web display, affiliate and review sites, publishers

Agents make the purchase step easy to move, because a buyer can reach any merchant through an agent. That makes the two things agents cannot easily copy more valuable: the place where preferences are formed, and the checkout where the sale is recorded. Meta is the only listed company building both at scale. Alphabet has YouTube for influence and Gemini and Shopping for the sale, but its core search business sits in the most exposed stage.

Five shifts to watch

From paying for attention to paying for the sale. If agents click less but buy more, advertisers will move from paying per impression or click towards paying a commission on each sale. That favours platforms that own checkout.
Fewer, more valuable ad slots. A handful of agents and stores will hold most of the ad space. Publishers, review sites and open-web ad technology firms, such as The Trade Desk, which analysts say faces "impression scarcity", lose the most.
The neutrality question. An agent that works for you but is paid to recommend things has a conflict of interest. Expect disclosure rules on sponsored recommendations, and a possible split between free, ad-funded agents and paid, ad-free ones. Apple is the obvious candidate to sell an agent that "works only for you", which would hurt Google's model most.
Marketing written for machines. Brands will spend more on clean product data, reviews and feeds that agents can read, and less on ads aimed at people in the research stage. Product feeds in Google Shopping and Shopify become more valuable.
A smaller research budget. If agents make comparison nearly free, merchants compete harder on price and need to buy less attention to win a sale. WPP notes that ad spending as a share of GDP is at its highest since its records began in 1999, which leaves room for that share to fall.

Four scenarios for advertising, 2027 to 2030

The odds are our rough judgement.

Scenario	Odds	What happens	Likely winners, such as	Likely losers, such as
1. Ads move inside the agent	40%	Sponsored answers become normal and the big platforms keep most of the money	Alphabet, Meta, Amazon	Publishers, affiliate and review sites, The Trade Desk
2. Commissions replace clicks	25%	Performance budgets move to fees on agent sales	Amazon, Walmart, Meta, Shopify	Alphabet's search margins, The Trade Desk, Criteo
3. Neutral agents win	15%	People prefer agents that take no ads	Apple, strong consumer brands	Alphabet most; the research ad budget shrinks
4. Slow change	20%	Agents stay mostly helpers, and search ads keep growing inside AI answers	Alphabet	Publishers, whose slide continues

1. Ads move inside the agent: 40% (base case)

Google, Meta, Amazon and OpenAI put sponsored results inside their agents and answers, and advertisers follow. Total ad spend keeps growing, but more of it is concentrated in fewer platforms.

Winners: Alphabet, if ads in AI Mode earn close to what classic search ads did; Meta; Amazon. OpenAI, which is private, becomes a large ad platform.
Losers: publishers and affiliate sites such as Ziff Davis, which has said Google's AI Overviews are cutting its search referrals, and open-web ad technology such as The Trade Desk.
How we'd know: Alphabet's Search revenue growth catches up with query growth; ad formats in AI Mode and ChatGPT spread to more countries; Google Network impressions keep falling.

2. Commissions replace clicks: 25%

Advertisers stop paying for clicks they cannot tie to sales and pay a fee on each sale instead. The money flows to whoever processes the purchase.

Winners: stores and checkout owners such as Amazon, Walmart and Shopify; Meta, if Muse commission fees take hold. TikTok Shop, which is private, gains too.
Losers: Alphabet's search margins, as cost per click comes under pressure; open-web ad technology such as The Trade Desk and Criteo; affiliate sites, as platforms keep the commission.
How we'd know: Meta reports its first Muse commission revenue; Amazon and Walmart ad growth stays well ahead of Google Search; YouTube discloses creator commission revenue.

3. Neutral agents win: 15%

Trust becomes the selling point. Enough people prefer agents that take no ads, paid for by subscription or bundled with a phone, that sponsored recommendations lose value.

Winners: Apple; strong consumer brands that people ask for by name; communities people trust, such as Reddit, if they can charge AI companies for their content.
Losers: Alphabet most; the research-stage ad budget shrinks; ad-funded agents.
How we'd know: Apple positions Siri as an ad-free agent; regulators require clear labels on paid recommendations; paid agent subscriptions grow faster than free ones.

4. Slow change: 20%

Agents stay mostly helpers, as in the base case of our consumer adoption post. Search ads inside AI Overviews keep growing, and the shift happens over a decade rather than a few years.

Winners: Alphabet, whose search business holds up longer.
Losers: publishers, whose traffic decline continues regardless.
How we'd know: Alphabet's Search growth re-accelerates as the AI Mode drag fades; Pinterest's ad pricing recovers; agent shopping stays a small share of Shopify's sales.

How each group fares across the scenarios

Group, such as	1. Ads inside the agent	2. Commissions	3. Neutral agents	4. Slow change
Alphabet	Gains	Loses margin	Loses most	Gains
Meta	Gains	Gains	Mixed	Gains
Amazon	Gains	Gains most	Mixed	Gains
Walmart	Gains	Gains	Mixed	Gains
Shopify	Mixed	Gains	Gains	Mixed
Pinterest	Mixed	Loses	Mixed	Mixed
Reddit	Mixed	Mixed	Gains	Loses traffic
The Trade Desk, Criteo	Loses	Loses	Loses	Loses slowly
Publishers and affiliate sites, such as Ziff Davis	Loses	Loses	Mixed	Loses slowly
Apple	Mixed	Mixed	Gains most	Mixed

The pattern is clear at the extremes. Amazon and Meta do well in most scenarios, because they combine influence or the store with closed-loop measurement. Open-web ad technology and publishers lose in all four. Alphabet has the widest range: it holds up if ads move smoothly into AI answers and loses most if agents turn neutral.

What this could mean for selling puts

This section is an AI-assisted view and is not financial advice. For any cash-secured put, the maximum loss is the strike minus the premium received, per share, if the stock goes to zero, and the breakeven is the same number.

Ad businesses report the shift through quarterly numbers such as cost per click, price per ad and user counts, so earnings dates carry most of the risk. Reddit's stock fell after a strong quarter because of its user warning, and Pinterest guided cautiously despite record users. Names in the "loses in all four" row can offer rich premium for a long time, but assignment there means owning a business in steady decline. The levelbox screener updates each name's trend and premium daily.

What to watch next, in order

Black Friday and Cyber Monday. TikTok Shop's US sales (last year's Black Friday weekend did about $500 million) and Shopify's agent orders.
The Trade Desk's Q3 report, due shortly, for the open web.
Amazon's Q4 ad growth against its high-teens guidance, and any Alexa for Shopping ad data.
Alphabet's Q4 report (early February). Search revenue against query growth, cost per click, Network impressions, and whether it keeps selling shares to fund capex.
Meta's Q4 report. The partnership ads run rate and any first Muse commission revenue.
Pinterest's Q4 ad pricing. A recovery would weaken the idea that influence without the sale gets squeezed.
Reddit's Q4 user numbers and the results of its logged-in test.

Method and sources

Market size and mix: WPP Media's June 2026 forecast and WARC's Q3 2026 forecast, as reported by exchange4media and PPC Land.
Alphabet: the Q3 2026 earnings release and Form 10-Q, the earnings call as summarised by Security Boulevard and 9to5Google, and Cleary Gottlieb on the June 2026 equity offerings.
Ziff Davis: its Q2 2026 earnings call (August 2026).
Publisher traffic: Chartbeat data via Search Engine Journal, Technology.org and Search Engine Land.
Reddit: the Q3 2026 shareholder letter filed with the SEC, and the earnings call.
Meta: the Q3 2026 earnings call transcript (Motley Fool, StockAnalysis).
Pinterest: the Q3 2026 press release filed with the SEC and the earnings call transcript.
YouTube: Alphabet's Q3 2026 results and Nielsen data via PPC Land.
TikTok Shop: Momentum Works and Tabcut estimates via Tubefilter. TikTok is private and does not report these figures.
Amazon and Walmart: Amazon's Q3 2026 results via Relevant Audience, CNBC and The Desk; Walmart's Q2 FY27 results via eMarketer and Yahoo Finance.
Creator spend: IAB estimates via Digiday and other secondary reports. We could not access the IAB report directly.
ChatGPT ads: OpenAI's advertiser figures via Retailgentic; click-rate estimates from independent benchmark reports, which vary.
Framework, scenario odds and winners and losers: our judgement, made with AI assistance. They are not model outputs.

Links:

This article is analytical and educational and is not financial advice. It names companies as worked examples only. Figures are a point-in-time snapshot as of 3 November 2026 and can contain errors. This article was generated with AI assistance. If you spot a mistake, tell us.

Common questions

How will AI agents change online advertising?: AI agents take over the research step of shopping: comparing products, reading reviews and checking prices. That is where search ads, review sites, affiliate links and open-web display ads earn most of their money, so those are the most exposed. Advertising that shapes preferences, such as creator content and video feeds, and advertising at the point of sale, such as sponsored listings inside a store, look likely to hold up or gain. Pricing is also likely to move from paying per click towards paying per sale.
Is Google search advertising declining?: Not yet in total. Alphabet's Search revenue grew 12% year on year in Q3 2026 to $63.35 billion. But it was flat from the previous quarter while search queries grew at their fastest pace in five years, which means each search is earning less. Management said lower monetisation of AI Mode, where queries run about three times longer, weighed on growth. Paid clicks rose 8% and cost per click 4%. Ads that Google places on other websites fell 14% in impressions.
Which companies benefit from AI agents in advertising?: In our AI-assisted view, the strongest position belongs to companies that both shape what people want and own the sale. Meta grew ad revenue 26% in Q3 2026 with price per ad up 14%, and is adding commission fees through its Muse agent. Retail media owners such as Amazon and Walmart are also well placed. Platforms that influence people but send them elsewhere to buy, such as Pinterest, are being pushed towards cheaper performance pricing. The names are worked examples only.
What is happening to publishers and Reddit's traffic from Google?: Chartbeat data shows Google Search referrals to publishers fell about 40% in the year to July 2026, while AI chatbots sent just 0.01% of publisher traffic. Reddit's US logged-out daily users fell 4% in Q3 2026, and it warned total daily users could fall by a low double-digit percentage in Q4 as Google's AI answers send fewer visitors. Reddit's revenue still grew 53%, because most of its money comes from logged-in users.
What does owning the sale mean in advertising?: A platform owns the sale when the purchase happens on its own service, so it can see which ads led to which sales and charge for results. Amazon, Walmart and TikTok Shop own the sale through their stores and checkout. Meta is building towards it with conversion measurement and commission fees on purchases made through Muse. Platforms that send shoppers to another website to buy have less proof of results and less pricing power.