How each claim is supported:
Tap any marked phrase for its source.
Section 1 of 9
How this industry makes money
In one sentence
A platform or publisher builds an audience, then earns from it. For many, advertising is almost everything: Meta earned $59.4 billion of its $60.8 billion from ads last quarter.
Six pools of money decide the year.
| Pool | What it is | Who controls it |
|---|---|---|
| Audience | Visitors and users, and where they come from | Search engines, app stores, social apps and the product itself |
| Advertising | Ad impressions times the price per ad | Advertisers, and the platforms they compare you to |
| Subscriptions and fees | What users pay directly | The business, within what users accept |
| Content cost | Staff, creators and licensed content | The business |
| Trust and safety | Moderation, fraud checks and legal compliance | The business, within the law |
| Computing | Servers, cloud and AI services | Cloud and AI providers |
AI touches all six. It changes where the audience comes from, how well ads are targeted, what content costs to make, how much moderation is needed, and how much computing costs.
Section 2 of 9
Where AI touches the economics
In one sentence
Audience: less from search. People who see an AI summary click through to websites about half as often. Businesses that rely on search traffic lose visitors, even when their content is used in the summary.
Advertising: better targeting wins. Meta's ad prices rose 12% as its AI tools improved results for advertisers. Advertisers move money to whoever proves the best return. Smaller platforms must show results in the same terms.
Content cost: lower. AI can draft, edit, translate and make images cheaply. A small team can publish far more.
Trust and safety: higher. Cheap AI content brings more spam, fakes and fraud to sort. In Europe, deepfakes must now be labeled and chatbots must say they are AI.
Computing: higher. AI features run on costly computing. Meta's costs rose partly because of AI infrastructure and fees paid to other AI companies.
Time saved is not money saved. Faster content creation pays only if the extra content earns more than it costs.
Section 3 of 9
What is already happening
In one sentence
| Who | What they did | Reported result | What the source does not show |
|---|---|---|---|
| Google search users (Pew, 2025) | Saw AI summaries on about one in five searches | Clicked a website link in 8% of visits, against 15% without a summary | Results since March 2025; Google disputes the method |
| Meta (2026) | Used AI to rank ads and help advertisers make them | Ad revenue up 27%; price per ad up 12% | How much of the gain came from AI alone |
| Meta (2026) | Spent heavily on AI infrastructure | Operating margin fell from 43% to 31% | When the spending pays back |
| Meta advertisers (2026) | Used Meta's AI tools to make ads | More than 9 million small businesses; clicks and conversions up on Facebook, by the company's account | Independent measurement of results |
| The European Union (2026) | Began applying AI transparency rules | Chatbot disclosure and deepfake labeling required from August 2, 2026 | How strictly they will be enforced |
Two patterns stand out. AI is making the largest platform stronger at selling ads, at a high cost. And businesses that depend on search for visitors are receiving fewer of them.
Section 4 of 9
What could change next
In one sentence
Scenario 1: Search sends less and less. AI answers appear on more searches, and fewer people click through. Assumes: search engines show AI summaries more often. Signal to watch: your own share of visitors from search, month by month. Last checked: September 2026. The best public study found clicks roughly halved where summaries appear.
Scenario 2: Ad money concentrates further. The largest platforms' AI targeting pulls more budget from smaller ones. Assumes: their AI keeps improving advertisers' results. Signal to watch: the largest platforms' ad price growth, and your advertisers' budgets shifting. Last checked: September 2026. Meta's price per ad rose 12% last quarter.
Scenario 3: Europe enforces its labeling rules. Regulators begin fining firms that fail to label AI content or disclose chatbots. Assumes: authorities choose early cases to set an example. Signal to watch: the first published enforcement action under the transparency rules. Last checked: September 2026. The rules apply from August 2, 2026, with fines up to €15 million or 3% of revenue.
All three reward the same preparation: a direct relationship with your audience, proof that your ads work, and a list of every AI feature that touches users in Europe.
Section 5 of 9
Who gains and who loses
In one sentence
Gain: the largest platforms. Their AI improves ad results, and ad money follows.
Gain: small advertisers. AI makes ads cheaper and faster to produce.
Gain: businesses with direct audiences. Email lists, apps and subscribers do not depend on search.
Lose: sites that live on search traffic. AI summaries answer the question on the search page.
Lose: smaller ad platforms without strong targeting. Advertisers compare results and move budgets.
Mixed: creators. AI makes content cheaper to produce, but it also floods platforms with competition.
Section 6 of 9
Where to look before you invest
In one sentence
| Where | Prerequisite | Cost range for a business with $20 million to $500 million in revenue | Most common reason it fails | Pitch you will hear |
|---|---|---|---|---|
| Direct audience: email, apps, subscriptions | A reason for people to come back without searching | Staff time and product work | Built as a side project nobody owns | "Own your audience" |
| Proof that ads work | Measurement advertisers trust, in their terms | $50,000 to $300,000 a year | Reporting clicks while advertisers judge sales | "AI-powered ad targeting" |
| AI for content and moderation | Editors who check what AI writes, and rules for what it may do | Varies with volume | Cheap content that damages trust | "Publish ten times more" |
| AI rules compliance | A list of every AI feature that touches users | Legal and product time | Nobody owns the list | "Fully compliant AI" |
The cost ranges are illustrative and will be replaced with reviewed figures. The order matters more than the amounts: an audience you reach directly is worth more than any tool that depends on someone else's platform.
One test before any contract. Ask the vendor to run its tool on last quarter's data and show the result against what actually happened.
Section 7 of 9
Up and down the chain
In one sentence
Game makers. Games compete for the same attention and ad budgets. See the guide for games.
Advertisers. Consumer brands and retailers are moving budgets toward the platforms that prove results.
Restaurants and local businesses. Local businesses use platforms to be found and booked.
Investors. Funds are repricing companies based on what AI will do to them.
Section 8 of 9
Questions for your team
The same ground, for whoever owns each answer. Bring them to your next leadership meeting.
Ask leadership
What share of visitors comes from search, and what is the trend?
A good answer
A monthly figure and a plan if it falls.
Push back if you hear
“Search has always worked.”
Ask head of revenue
Can advertisers see results in the terms they judge other platforms by?
A good answer
Measurement they use to decide budgets.
Push back if you hear
“Clicks are strong.”
Ask head of product
Which features use AI, and which touch users in Europe?
A good answer
A current list with an owner for each.
Push back if you hear
“Legal is looking at it.”
Ask head of content
Who checks what AI writes before it goes out?
A good answer
Named editors and rules.
Push back if you hear
“The tool is accurate.”
Ask finance
What does AI computing cost per user, and is it rising?
A good answer
A figure by feature.
Push back if you hear
“It is in the cloud bill.”
Ask general counsel
Which chatbots and AI images must be labeled under the new rules, and are they?
A good answer
A checked list, reviewed each quarter.
Push back if you hear
“The rules were delayed.”
Section 9 of 9
What changed
In one sentence
September 29, 2026: first published.
August 2026 (reflected at launch): Europe's AI transparency rules began to apply. Affects sections 2, 3, 4 and 6.
July 2026 (reflected at launch): Meta reported 27% ad revenue growth and a lower margin. Affects sections 1, 3 and 5.
July 2025 (reflected at launch): Pew found that AI summaries roughly halve clicks to websites. Affects sections 1 through 4, and sets the guide's main conclusion: reduce dependence on search.
Decision exercise
Write your prediction and one line of reasoning before you compare it with the guide.
Sources
Pew Research Center, Google users are less likely to click on links when an AI summary appears in the resultsJuly 22, 2025
The Register, Google AI Overviews are killing the web, Pew study shows (again)July 22, 2025
Meta, Meta reports second quarter 2026 resultsJuly 29, 2026
Meta, Form 10-Q for the second quarter of 20262026
Adgully, Meta Q2 ad revenue rises 27% as AI tools boost advertiser growthJuly 2026
Ropes & Gray, Operationalising the EU AI Act's transparency obligationsAugust 2026
Compliance Hub, EU AI Act Article 50 transparency and the Digital Omnibus2026