AI Answer Engines and Zero-Click Ecosystem Dominance

1. Introduction

AI answer engines are systems that provide a synthesized answer directly to a user's question instead of merely presenting a list of links.

Traditional search:

User → Search engine → Links → User visits websites

AI answer engine:

User → AI → Synthesized answer → User may never visit underlying websites

This produces the phenomenon commonly called a zero-click ecosystem.

A zero-click ecosystem is an environment in which the user obtains the information, recommendation, comparison, or even transaction without clicking through to the underlying content provider.

From a competition-law perspective, the important question is:

Can a dominant AI answer engine use its position as the user's information gateway to control traffic, data, recommendations, advertising, transactions, or access to competing services?

There is no established EU legal category called "AI answer-engine dominance" or "zero-click dominance." The analysis must instead use existing principles concerning dominance, self-preferencing, foreclosure, tying, interoperability, essential inputs, data advantages, and platform ecosystems.

2. What Is an AI Answer Engine?

An AI answer engine combines several functions:

search;

retrieval;

ranking;

summarisation;

reasoning;

source selection;

recommendation;

personalization;

sometimes transaction execution.

For example:

Traditional search

User asks:

"Best hotels in Paris."

The search engine provides:

Hotel A

Hotel B

Hotel C

Booking website

Travel website

Reviews

The user clicks different websites.

AI answer engine

The AI says:

"For a three-night stay, Hotel A is suitable because..."

The user may not visit:

Hotel A's website;

travel comparison websites;

review websites;

independent publishers.

This changes the competitive structure.

3. Meaning of Zero-Click Economics

A zero-click system effectively compresses:

Multiple websites → AI retrieval → One answer

The user may therefore stop at the AI interface.

This can shift economic value from:

Content providers

to:

AI intermediary

because the intermediary controls the user's attention.

4. The Zero-Click Value Chain

A simplified structure is:

Publisher / Merchant / Service Provider

↓

Search / Data / Content

↓

AI Answer Engine

↓

User

↓

Purchase / Decision / Action

The AI platform may therefore sit between information suppliers and consumers.

If the platform becomes dominant, it could potentially control:

which information is retrieved;

which sources are cited;

which sources receive traffic;

which products are recommended;

which services are called;

which transactions are completed.

5. Why This Can Create Competition Concerns

The central issue is intermediation power.

A website historically competed for:

clicks

An AI answer engine may compete for:

the user's entire decision process.

Therefore, the platform could become a decision gateway.

This creates several possible competition concerns:

self-preferencing;

traffic foreclosure;

source discrimination;

data extraction;

tying;

interoperability restrictions;

vertical integration;

advertising conflicts;

exclusion of specialist search providers;

leveraging search power into downstream markets.

6. Case Law 1 — Google Shopping

Google and Alphabet v Commission

Case C-48/22 P
CJEU, 10 September 2024

This is the most important existing EU authority by analogy.

Google's general search service displayed its own comparison-shopping service more prominently while competing comparison-shopping services were generally demoted. The CJEU dismissed Google's appeal and upheld the General Court's judgment concerning the Article 102 infringement. (curia)

Principle

A dominant platform's method of displaying its own service can constitute exclusionary conduct where it gives the affiliated service preferential treatment capable of restricting competition.

AI answer-engine application

Imagine:

AI Answer Engine

User:

"Compare smartphones under €500."

The AI has access to:

independent technology websites;

comparison services;

manufacturer websites;

its own shopping service.

If the AI systematically gives its own shopping service superior exposure, the Google Shopping reasoning becomes highly relevant by analogy.

Important distinction

AI summarisation itself is not unlawful.

The question is whether the platform's method of selecting and presenting information uses dominance to exclude competing services.

7. From Search Ranking to Answer Ranking

Google Shopping involved:

Which shopping results appear prominently?

AI answer engines may create a new question:

Which information becomes part of the answer at all?

This is potentially more powerful.

Traditional ranking:

Result 1
Result 2
Result 3
Result 4

AI answer:

"The answer is X."

The user may never know that alternative sources existed.

Therefore:

Search ranking

controls visibility.

AI answer ranking

can potentially control inclusion in the answer itself.

8. Zero-Click Traffic Foreclosure

Suppose a news publisher receives:

1 million monthly visits from search.

An AI answer engine begins answering users directly.

Traffic becomes:

100,000 visits.

If the AI simultaneously promotes its own content or services, the affected publisher may face:

lower advertising revenue;

fewer subscriptions;

reduced brand recognition;

reduced ability to invest in content;

loss of data about users.

This could theoretically create traffic foreclosure.

However:

A decline in traffic caused by technological substitution is not automatically an Article 102 infringement.

The authority would need to establish dominance, abusive conduct, and the required exclusionary effects or capability.

9. Case Law 2 — Microsoft

Microsoft Corp. v Commission

Case T-201/04
General Court, 17 September 2007

Microsoft concerned, among other issues:

refusal to supply interoperability information;

interoperability;

tying Windows with Windows Media Player.

The General Court upheld important parts of the Commission's Article 82 decision. (Infocuria)

AI relevance

Suppose an AI answer engine controls the interface through which users access:

search;

browsers;

publishers;

shopping services;

external AI tools.

Independent providers may need APIs or technical interoperability to participate effectively.

A dominant answer engine might theoretically restrict:

API access;

source integration;

indexing;

external tool connections;

citation interfaces.

Microsoft provides an important analogy for analyzing whether technological control is being used to disadvantage interoperable competitors.

10. Zero-Click and Interoperability

Consider:

Publisher A

↓

wants AI citation traffic

↓

AI platform API

↓

access denied

Meanwhile:

Platform's own content

↓

fully integrated

↓

AI answer

This creates a possible input-access and interoperability issue.

The key legal question would be whether the external access is genuinely necessary and whether the conduct meets the demanding conditions established by EU refusal-to-supply jurisprudence.

11. Case Law 3 — Bronner

Oscar Bronner GmbH & Co. KG v Mediaprint

Case C-7/97
CJEU, 26 November 1998

Bronner concerned access to a dominant undertaking's newspaper home-delivery system.

The CJEU applied a strict test concerning refusal to provide access to infrastructure developed for the dominant undertaking's own business. The infrastructure had to be indispensable, among other requirements. (curia)

AI relevance

Suppose a publisher argues:

"The dominant AI answer engine is the only way consumers can discover my content."

That statement would not automatically establish a legal right to access.

The analysis would examine:

indispensability;

alternative routes to users;

actual/potential substitutes;

elimination of competition;

objective justification.

Thus:

Important distribution channel ≠ automatically indispensable facility.

12. Case Law 4 — IMS Health

IMS Health GmbH & Co. OHG v NDC Health GmbH & Co. KG

Case C-418/01
CJEU, 29 April 2004

IMS Health concerned a specialized pharmaceutical data structure and a refusal to license access.

Principle

The case developed the exceptional circumstances under which refusal to license/access a protected resource can amount to abuse of dominance.

AI application

Consider a dominant AI answer engine with a unique:

knowledge graph;

content index;

structured database;

user-intent dataset;

retrieval infrastructure.

An independent answer engine might claim:

"Without access to this dataset, I cannot compete."

IMS Health demonstrates that high commercial value is not automatically equivalent to legal indispensability.

13. The Data Problem

AI answer engines can accumulate enormous quantities of:

user queries;

clicks;

search history;

corrections;

feedback;

source performance;

product preferences;

transaction outcomes.

This can create:

More users → More data → Better AI → More users

This is a data-network feedback loop.

It may create entry barriers for smaller answer engines.

But again:

A data advantage resulting from successful competition is not automatically unlawful.

The legal question is whether the undertaking uses that advantage through conduct prohibited by competition law.

14. Case Law 5 — Meta Platforms

Meta Platforms Ireland Ltd and Others v Bundeskartellamt

Case C-252/21
CJEU, 4 July 2023

Meta combined data from Facebook and other sources, including activities on third-party websites and applications. The CJEU held that a competition authority examining abuse of dominance could consider GDPR compliance, while respecting the competence of data-protection authorities. (curia)

AI relevance

An AI answer engine may combine:

search queries;

browsing;

purchases;

location;

third-party applications;

user interactions.

The resulting data concentration may strengthen the platform's position.

Meta demonstrates that competition law and data governance can interact.

Important qualification

A GDPR violation does not automatically establish an Article 102 infringement.

The competition analysis still requires the relevant legal elements.

15. Zero-Click and Publisher Data

Traditional websites receive valuable information when users click:

search terms;

referrer;

browsing behaviour;

session duration;

conversion;

return visits.

With zero-click answers, much of that interaction may remain inside the AI platform.

Therefore:

Publisher → loses direct user relationship

while:

AI platform → gains user interaction data

This can reinforce the platform's data advantage.

16. Case Law 6 — Google Android

Google and Alphabet v Commission

Case T-604/18
General Court, 14 September 2022

The case concerned Google's Android ecosystem, including:

Android operating systems;

Google Play Store;

Google Search;

Chrome;

agreements with device manufacturers and mobile network operators;

product bundles;

exclusivity payments;

anti-fragmentation obligations.

The General Court characterized the dispute in terms of a multi-sided platform/ecosystem and exclusionary effects. (Infocuria)

AI relevance

An AI answer engine may similarly operate as an ecosystem:

AI

↓

Search

↓

Browser

↓

Advertising

↓

Shopping

↓

Applications

↓

Payments

The platform could potentially leverage power between these interconnected markets.

Key concept

AI answer engine ≠ isolated product.

It may become an ecosystem.

17. Bundling and Tying

An AI platform could potentially say:

"To receive our premium AI answers, use our browser."

or:

"Our AI answer engine works fully only with our search index."

or:

"External shopping services receive reduced functionality."

This can raise questions analogous to Microsoft and Google Android.

The legal analysis depends upon:

market definition;

dominance;

separate products/services;

coercion or contractual structure;

foreclosure;

efficiencies;

competitive effects.

18. Case Law 7 — Eturas

Eturas UAB and Others v Lietuvos Respublikos konkurencijos taryba

Case C-74/14
CJEU, 21 January 2016

Travel agencies used a common computerized booking system. The system administrator imposed an automatic restriction on online discounts and sent a system message concerning that restriction. The CJEU examined whether this could establish a concerted practice and addressed evidentiary questions. (Infocuria)

AI relevance

An AI answer engine could theoretically become a common technological intermediary used by many businesses.

For example:

Multiple retailers

↓

Common AI platform

↓

AI recommends similar prices/products

If the system facilitates coordination, Article 101 questions could arise.

But:

Similar AI-generated outcomes do not automatically prove a cartel.

Evidence of the required agreement or concerted practice remains necessary.

19. Zero-Click and Article 101

Article 101 issues could arise if an answer engine facilitates:

coordinated pricing;

exchange of competitively sensitive information;

market allocation;

common commercial restrictions.

For example:

Retailer A

Retailer B

Retailer C

↓

common AI system

↓

access to future pricing information

↓

coordinated conduct.

The technological intermediary does not eliminate the need to establish the legal elements of Article 101.

20. Case Law 8 — Google AdSense

Google and Alphabet v Commission — Google AdSense

Case concerning Google's AdSense conduct

The EU Google advertising cases are relevant because advertising intermediation can give a platform control over how publishers monetize traffic.

AI relevance

A zero-click AI platform could potentially control both:

information discovery

and

advertising monetisation.

That creates a potentially important conflict:

The same platform may determine which information users receive and which commercial messages accompany that information.

The competition analysis would examine the relevant markets and specific exclusionary practices rather than assuming that vertical integration is unlawful.

21. The Zero-Click Advertising Problem

Traditional model:

Publisher

↓

User visits publisher

↓

Advertisement

↓

Publisher earns revenue.

Zero-click model:

User

↓

AI answer

↓

No publisher visit

↓

Potentially AI-controlled advertising.

Therefore, the AI platform may capture a larger share of the economic value generated by information discovery.

Possible concerns include:

publisher foreclosure;

advertising-market leveraging;

data concentration;

vertical integration;

self-preferencing.

22. Answer Engines as Multi-Sided Platforms

An AI answer engine can serve:

Users

Want accurate answers.

Publishers

Want visibility and traffic.

Advertisers

Want consumers.

Merchants

Want sales.

Developers

Want distribution.

Data providers

Want commercial relationships.

AI models

Want compute and users.

The platform coordinates all these sides.

This can create cross-side network effects.

23. Zero-Click Network Effects

The cycle may be:

Users ↑

↓

Queries ↑

↓

Data ↑

↓

AI quality ↑

↓

More users ↑

At the same time:

Users ↑

↓

Advertisers ↑

↓

Revenue ↑

↓

AI investment ↑

↓

AI quality ↑

This can make entry increasingly difficult.

24. The Publisher Dependency Problem

Publishers may become dependent on AI systems for:

discovery;

traffic;

citations;

reputation;

subscriptions.

If the AI changes its algorithm, the publisher's traffic could fall.

This creates a potential gatekeeper relationship.

But the legal analysis must distinguish:

Normal technological competition

A new technology changes consumer behaviour.

from:

Exclusionary conduct

A dominant undertaking deliberately uses its position to disadvantage rivals.

25. Zero-Click Self-Preferencing

Consider:

User

"Which financial product should I choose?"

AI

The platform's own financial comparison service is used internally.

Independent comparison sites are not displayed.

Potential concerns:

self-preferencing;

foreclosure;

discrimination;

lack of interoperability.

Google Shopping provides the closest major EU competition-law analogy.

26. Citation Manipulation

An AI answer engine might determine:

which source is cited;

how frequently it is cited;

whether a source is cited at all.

Potential competition questions include:

Does the AI give its own content systematic citation advantages?

Does it reduce visibility of competing information providers?

Does it use commercial relationships to influence source selection?

Again, the existence of a ranking algorithm alone does not establish an infringement.

27. Content Extraction vs Traffic Generation

AI answer engines may face a structural tension.

Traditional search

Content → Search result → Click → Publisher

AI answer engine

Content → AI → Answer

The AI may derive substantial value from the underlying information without sending the user to the original provider.

This creates a potential zero-click extraction model.

28. Is Zero-Click Itself Anticompetitive?

No.

This distinction is essential.

Zero-click functionality can be an innovation that:

saves time;

reduces search costs;

improves accessibility;

makes information easier to understand.

Competition law does not require platforms to preserve outdated click-based business models.

The issue becomes more serious when:

dominant position + exclusionary conduct + foreclosure/effects

are established.

29. The Bronner Problem for Publishers

A publisher might argue:

"The AI platform must send users to my website."

That argument is not automatically successful.

Bronner demonstrates the demanding nature of compulsory-access claims.

Questions include:

Is AI access indispensable?

Are there alternatives?

Can the publisher reach users elsewhere?

Would refusal eliminate competition?

Is the refusal objectively justified?

Therefore:

Loss of traffic ≠ automatic right to platform access.

30. The AI Knowledge Graph as a Potential Input

AI answer engines may build knowledge structures from:

public information;

licensed information;

proprietary data;

user-generated information.

If the platform controls an especially important information infrastructure, competitors might seek access.

This creates an analogy with:

IMS Health

Magill

Bronner

But compulsory-access doctrines remain exceptional.

31. Vertical Integration

Imagine a company owns:

Search

AI Answer Engine

News Service

Shopping

Advertising

Cloud

Payments

The AI answer engine can become the central distribution layer.

It could theoretically direct:

information;

customers;

advertising;

transactions;

payments.

This creates opportunities for leveraging.

32. Raising Rivals' Costs

A dominant AI platform might potentially make access expensive through:

API charges;

indexing fees;

data licensing;

certification;

ranking fees;

technical requirements.

Independent services might remain formally available but become commercially difficult to operate.

This can raise a potential raising-rivals'-costs theory.

33. Margin Squeeze

Suppose an AI platform controls an upstream information-access service.

It charges:

Independent competitor → high access fee

while its own downstream service receives:

Internal access → low/no fee

The independent competitor may then be unable to compete.

This creates a possible margin-squeeze theory.

The precise legal requirements depend on the market structure and applicable jurisprudence.

34. Free AI and Predatory Concerns

Many answer engines may be offered for free.

That does not automatically mean predatory pricing.

The platform might monetize through:

advertising;

subscriptions;

cloud services;

enterprise contracts;

commissions;

transactions.

Competition authorities may therefore need to examine the whole ecosystem rather than simply looking at the consumer price.

35. Data Feedback and Entry Barriers

A successful answer engine can learn from:

user questions;

corrections;

clicks;

source quality;

user satisfaction;

transaction results.

New entrant:

Small user base → limited data → weaker optimization

Dominant platform:

Large user base → massive data → better optimization

This can create an endogenous entry barrier.

But:

Data advantage is not automatically abusive.

The legal issue concerns the means by which market power is acquired, maintained or leveraged.

36. Consumer Choice

Potential competitive harms could include:

fewer sources;

fewer competitors;

higher prices;

lower quality;

less innovation;

reduced diversity;

reduced privacy.

But AI answer engines can simultaneously produce:

lower search costs;

faster decisions;

easier comparison;

improved accessibility;

better personalization.

Thus an investigation should examine both competitive harm and efficiencies.

37. Evidence in a Zero-Click Investigation

A competition authority might examine:

Traffic data

clicks;

referrals;

impressions;

conversion rates.

AI output data

source selection;

citation frequency;

rankings;

recommendation patterns.

Algorithmic evidence

system prompts;

ranking rules;

retrieval policies;

source-weighting mechanisms.

Commercial evidence

contracts;

advertising arrangements;

exclusivity;

payments.

Data evidence

query data;

click data;

user profiles;

source-performance information.

Economic evidence

foreclosure;

market shares;

switching costs;

entry barriers;

counterfactual outcomes.

38. Counterfactual Analysis

Suppose the AI platform's own service receives:

80% of recommendations.

That number alone does not prove abuse.

The relevant question might be:

What percentage would the platform's service receive if the AI ranked competing services according to a neutral quality-based criterion?

Possible comparison:

Actual

Platform service: 80%

Counterfactual

Platform service: 30%

If robust evidence supports such a counterfactual, it may help assess the competitive significance of the ranking mechanism.

39. Relevant Markets

Potential markets could include:

general search;

AI answer services;

specialized search;

online advertising;

content distribution;

shopping comparison;

travel comparison;

news discovery;

digital assistants;

AI agent marketplaces.

Market definition will depend on substitutability and the facts.

An AI answer engine could potentially compete simultaneously across several markets.

40. Important Case-Law Table

41. Main Competition Theories

1. Self-preferencing

AI favors its own services.

2. Traffic foreclosure

AI prevents users from reaching competing websites.

3. Data foreclosure

AI accumulates data that competitors cannot obtain.

4. Tying

AI functionality is tied to another service.

5. Bundling

Search + AI + shopping + advertising are integrated.

6. Interoperability restrictions

Competitors cannot effectively integrate.

7. Exclusive dealing

Publishers or service providers are restricted from using alternatives.

8. Leveraging

AI/search dominance is extended into another market.

9. Algorithmic coordination

AI infrastructure facilitates coordination between competitors.

42. Zero-Click Ecosystem Formula

Zero-Click Dominance

AI Answer Engine

Large User Base

Data Advantage

Network Effects

Control of Ranking/Source Selection

Vertical Integration

↓

Information & Distribution Bottleneck

↓

Potential Traffic/Data Foreclosure

↓

Self-Preferencing / Tying / Interoperability Restrictions / Leveraging

↓

Potential Article 102 TFEU Concern

43. Most Important Legal Distinction

The following proposition should not be assumed:

"AI answers eliminate website clicks, therefore AI is abusing dominance."

That would be too broad.

The proper legal sequence is:

Relevant market

↓

Dominance

↓

Specific conduct

↓

Actual or potential exclusionary effect

↓

Causal connection

↓

Objective justification / efficiencies

↓

Legal assessment

Recent EU jurisprudence emphasizes that Article 102 analysis generally requires identifying conduct other than competition on the merits that has actual or potential effects capable of restricting competition; actual successful exclusion is not always required. (Infocuria)

44. Six Cases to Memorize

1. Google Shopping — C-48/22 P

Keyword: Self-preferencing.

2. Microsoft — T-201/04

Keyword: Interoperability.

3. Bronner — C-7/97

Keyword: Indispensability.

4. IMS Health — C-418/01

Keyword: Critical information infrastructure.

5. Google Android — T-604/18

Keyword: Ecosystem and tying.

6. Meta Platforms — C-252/21

Keyword: Data + dominance.

Two additional cases are especially useful:

Eturas — C-74/14 → computerized coordination

T-Mobile Netherlands — C-8/08 → information exchange

45. Ultra-Simple Exam Explanation

AI answer engine = AI gives the answer instead of only giving links.

Zero-click = user gets what they need without visiting the underlying website.

Competition concern = if a dominant AI platform controls the gateway to users, it may potentially control:

traffic;

rankings;

data;

recommendations;

advertising;

transactions.

If that control is used to exclude competitors, potential issues include:

Self-preferencing + foreclosure + tying + interoperability restrictions + data leveraging + Article 102.

Conclusion

AI answer engines can fundamentally change the economics of digital intermediation.

The traditional internet model was:

Search → Click → Website

The emerging model can be:

Question → AI Answer → Action

The AI therefore potentially becomes not merely a search intermediary, but a distribution, recommendation and decision intermediary.

This creates a potential competitive bottleneck because publishers, merchants, comparison services, advertisers and other digital providers may increasingly depend on the AI platform for access to users.

The most relevant existing authorities are Google Shopping for preferential treatment in digital ranking, Microsoft for interoperability and tying, Bronner and IMS Health for access to potentially indispensable infrastructure or information, Google Android for ecosystem leveraging, Meta for data-related competition questions, and Eturas for computerized systems and competition-law coordination. (curia)

One-line exam formula:

AI Answer Engine Dominance = User Gateway + Zero-Click Intermediation + Data/Network Effects + Ranking Control + Vertical Integration → Potential Traffic/Data Foreclosure → Self-Preferencing, Tying or Leveraging → Possible Competition-Law Concern, subject to proof of dominance, abusive conduct and competitive effects.