Competition Law and Competition Intelligence Systems for Regulators

1. Introduction

Competition Intelligence Systems (CIS) are technological and analytical systems used by competition authorities to collect, integrate, monitor, analyse and interpret market information in order to identify possible anti-competitive conduct.

The concept is not, by itself, a separate offence under competition law. Rather, it represents an enforcement infrastructure that can help regulators detect and investigate:

cartels;

bid-rigging;

price coordination;

abuse of dominance;

exclusionary conduct;

discriminatory practices;

suspicious mergers;

digital-platform conduct;

algorithmic coordination;

market concentration;

excessive market power;

consumer harm.

Modern competition authorities increasingly need such systems because markets generate enormous quantities of digital information. The European Commission, for example, maintains searchable databases of competition cases and conducts sector inquiries, while its Article 102 procedures permit authorities to obtain information necessary for investigations. (Competition Policy)

2. Meaning of Competition Intelligence Systems

A Competition Intelligence System for Regulators may be defined as:

An integrated technological and analytical framework through which a competition authority collects, processes, analyses and visualises market information to detect, assess, investigate and monitor potential violations of competition law.

The system may combine:

economic data;

company information;

transaction data;

pricing information;

procurement records;

market shares;

consumer complaints;

corporate filings;

websites;

public databases;

investigation records;

communications obtained lawfully during investigations;

algorithmic outputs;

merger notifications.

The objective is not simply to collect data.

The objective is:

Data → intelligence → competition hypothesis → investigation → legal assessment → enforcement/remedy

3. Why Regulators Need Competition Intelligence Systems

Traditional competition enforcement was heavily dependent on:

complaints;

whistle-blowers;

inspections;

documentary evidence;

interviews;

market studies.

Modern markets create additional difficulties.

For example:

Digital markets

Prices may change every few seconds.

E-commerce

Millions of transactions may occur daily.

Algorithms

Pricing decisions can be automated.

Platforms

A single platform may operate several interconnected markets.

Cartels

Communication can occur through encrypted or digital channels.

Mergers

The competitive significance of a target may depend on data, users and future innovation rather than current turnover.

Consequently, competition authorities increasingly require analytical systems capable of identifying unusual patterns before deciding whether formal investigation is justified.

The CMA has specifically described the growing importance of algorithms and developed work aimed at identifying competition and consumer harms arising from algorithmic systems. (GOV.UK)

4. Main Components of a Competition Intelligence System

A sophisticated CIS can contain several layers.

A. Data collection layer

Collects information from:

merger filings;

company reports;

public procurement;

prices;

websites;

consumer complaints;

market surveys;

regulatory filings;

investigation documents.

B. Data integration layer

Combines information from different sources.

C. Analytical layer

Uses:

statistics;

econometrics;

network analysis;

machine learning;

anomaly detection;

market-share calculations.

D. Intelligence layer

Identifies:

suspicious patterns;

possible cartels;

emerging dominance;

unusual price movements;

merger risks.

E. Investigation layer

Provides investigators with:

leads;

timelines;

relationships;

documents;

evidence trails.

F. Monitoring layer

Checks whether firms comply with:

commitments;

behavioural remedies;

access obligations;

merger conditions.

5. Competition Intelligence vs Competition Investigation

These concepts should not be confused.

Competition intelligence

Primarily asks:

“Where might a competition problem exist?”

Competition investigation

Asks:

“Has a legally prohibited practice actually occurred?”

Therefore:

Intelligence = detection and prioritisation

Investigation = evidence gathering and legal determination

This distinction is extremely important because an algorithmic alert cannot itself establish a legal infringement.

6. Early-Warning Function

One of the most valuable functions of a CIS is early detection.

For example, suppose a regulator monitors 10,000 products.

The system detects:

simultaneous price increases;

identical price movements;

unusual timing;

reduced price dispersion.

This may produce a:

Cartel-risk alert

But the alert does not prove a cartel.

Investigators must determine whether the pattern resulted from:

lawful market conditions;

common costs;

supply shocks;

publicly available information;

unilateral optimisation;

actual coordination.

7. Cartel Detection

Cartels are particularly suitable for intelligence-based screening.

Possible indicators include:

parallel pricing;

bid rotation;

identical bids;

unusual winning patterns;

market allocation;

suspiciously stable market shares;

communication patterns;

simultaneous withdrawal of discounts;

abnormal tender participation.

The European Commission's cartel enforcement system includes published case information, investigation records and decisions, demonstrating the importance of structured competition intelligence and case databases. (Competition Policy)

8. Bid-Rigging Detection

Public procurement generates large datasets.

A regulator can analyse:

bidder participation;

winning frequency;

bid differences;

tender rotation;

geographic patterns;

subcontracting;

repeated combinations of bidders.

Example

Suppose:

Company A wins Tender 1

Company B wins Tender 2

Company C wins Tender 3

and the same three companies rotate winners over hundreds of tenders.

An intelligence system can flag the pattern.

Investigators can then examine whether there is evidence of:

bid rotation;

market allocation;

communication;

coordinated tendering.

9. Price Intelligence

A regulator can monitor prices across:

online retailers;

fuel stations;

airlines;

hotels;

pharmaceuticals;

food markets;

financial products.

The system can calculate:

Price Dispersion=Pmax−PminPrice\ Dispersion = P_{max}-P_{min}

and monitor:

ΔPrice=Pt−Pt−1\Delta Price = P_t-P_{t-1}

Sudden parallel movements may justify further analysis.

But parallel pricing is not automatically unlawful.

10. Algorithmic Competition Intelligence

Algorithms can be used by both:

Firms

to determine:

prices;

rankings;

discounts;

advertising;

recommendations.

Regulators

to detect:

suspicious patterns;

discriminatory outcomes;

exclusionary conduct;

algorithmic coordination.

The CMA has specifically noted that increasingly sophisticated algorithms can make competition harms harder to identify and has stated that its information-gathering powers can cover data, code and documentation needed to understand and test algorithmic systems. (GOV.UK)

11. Algorithmic Collusion

One major concern is whether algorithms can facilitate coordination.

Possible mechanisms include:

Explicit programming

Competitors intentionally design algorithms to coordinate.

Signalling

Algorithms react to competitors' prices.

Learning systems

Machine-learning systems independently discover stable pricing strategies.

Monitoring

Algorithms detect deviations from coordinated behaviour.

Competition intelligence systems can monitor these patterns.

However:

Algorithmic similarity is not automatically evidence of an unlawful agreement.

Human and economic investigation remains essential.

12. Digital Platforms

Digital platforms create unique competition-intelligence problems because they may simultaneously control:

data;

search;

advertising;

rankings;

payments;

marketplaces;

cloud infrastructure.

A regulator may therefore need to monitor several markets simultaneously.

The European Commission and U.S. competition authorities have recognized that digital enforcement increasingly requires analysis of network effects, large quantities of data and interoperability. (Federal Trade Commission)

13. Data as a Competition Asset

A CIS can analyse:

data concentration;

user growth;

switching rates;

multi-homing;

platform dependency;

access to datasets.

Data can create competitive advantages through:

Data → better prediction → better service → more users → more data

A regulator therefore needs intelligence capable of identifying such feedback loops.

14. Abuse of Dominance Detection

Competition intelligence can help regulators monitor dominant firms for:

predatory pricing;

discriminatory pricing;

tying;

bundling;

refusal to deal;

exclusive dealing;

self-preferencing;

discriminatory access;

margin squeeze.

Under EU law, for example, Article 102 investigations begin with market definition and assessment of dominance, after which authorities investigate potentially abusive conduct. (Competition Policy)

In India, similar analytical work supports inquiries under Section 4 of the Competition Act, 2002.

15. Merger Intelligence

CIS can monitor:

acquisition announcements;

investment patterns;

venture-capital transactions;

changes in ownership;

common directors;

minority investments;

technology acquisitions.

This is important because a transaction may be competitively significant even where traditional financial indicators are small.

The system can flag:

Dominant firm + rapidly growing start-up + high strategic value

for further merger analysis.

16. Monitoring Digital Ecosystems

A competition authority can construct an ecosystem map showing:

Platform

↓

Users

↓

Advertisers

↓

Developers

↓

Payment services

↓

Cloud services

↓

Data

This helps regulators understand whether one undertaking is gaining power across multiple connected markets.

17. Network Analysis

Network analysis is especially useful for cartel investigations.

Entities can be represented as:

Nodes = firms/individuals

Edges = communications/transactions/relationships

For example:

Company A ───── Company B    │              │    │              │ Company C ───── Company D

If the same individuals repeatedly connect competing firms during suspicious periods, investigators can prioritise those relationships for lawful examination.

The network itself is not proof of an infringement.

18. Consumer Complaint Intelligence

Competition regulators receive complaints from:

consumers;

suppliers;

competitors;

distributors;

employees;

industry associations.

A CIS can classify complaints according to:

sector;

company;

conduct;

geography;

frequency;

severity.

For example:

400 complaints concerning discriminatory access to a platform

could trigger a sector-level inquiry.

19. Sector Monitoring

Competition authorities can build sector dashboards for:

banking;

pharmaceuticals;

telecom;

aviation;

e-commerce;

energy;

food;

automobiles;

digital services.

The European Commission has used sector inquiries to gather extensive evidence. Its e-commerce inquiry, for example, gathered information from nearly 1,900 companies and analysed approximately 8,000 distribution contracts. (Competition Policy)

This demonstrates how large-scale information gathering can support competition-policy intelligence.

20. Competition Intelligence and Amazon

A particularly relevant example is the UK's investigation into Amazon's Marketplace.

The CMA examined:

third-party seller data;

Buy Box selection;

Prime delivery arrangements.

The CMA ultimately accepted commitments concerning Amazon's use of third-party seller data and Buy Box treatment. (GOV.UK)

This demonstrates how competition intelligence can combine:

platform data + algorithmic decision-making + commercial information + market analysis.

21. Competition Intelligence and Cloud Markets

The UK's cloud-services market investigation illustrates another intelligence function.

The CMA examined the structure of public cloud infrastructure and ultimately recommended prioritising strategic-market-status investigations involving the two largest providers for their respective digital activities. (GOV.UK)

This shows that intelligence systems can support market-wide structural monitoring, rather than merely individual infringement investigations.

22. Legal Limits on Competition Intelligence Systems

Competition intelligence must operate within legal constraints.

Important safeguards include:

1. Legality

Data must be collected under lawful authority.

2. Relevance

Information should be relevant to the competition inquiry.

3. Proportionality

Regulators should avoid excessive collection.

4. Confidentiality

Business secrets must be protected.

5. Due process

Companies must have an opportunity to respond to allegations.

6. Evidentiary reliability

Automated predictions should not automatically be treated as proof.

7. Human oversight

Important enforcement decisions should remain subject to accountable legal decision-making.

23. False Positives

A major danger is false positives.

Example:

A system detects:

Companies A, B and C increased prices simultaneously.

Possible explanations include:

cartel;

common increase in raw-material costs;

government tax increase;

supply shortage;

exchange-rate movement;

common public information.

Therefore:

Algorithmic alert ≠ legal conclusion.

24. False Negatives

The opposite problem is a false negative.

A sophisticated cartel may deliberately avoid obvious patterns.

For example:

irregular price changes;

different bid margins;

indirect communications;

rotating customers;

coded communications.

Therefore, a CIS should not become the only enforcement mechanism.

25. Explainability

Competition intelligence systems should be capable of explaining:

Why was this company or transaction flagged?

For example:

Alert score: high

because:

90% bid overlap;

repeated tender rotation;

abnormal bid differences;

common subcontractors;

suspicious timing.

An unexplained:

“AI says cartel”

is legally inadequate.

26. Human Oversight

Human experts should review automated alerts.

A suitable model is:

AI detection

↓

Economic analyst review

↓

Legal assessment

↓

Investigation

↓

Evidence verification

↓

Enforcement decision

This protects against automated errors.

27. Competition Intelligence and Confidential Business Information

Competition authorities often possess highly sensitive information.

Examples:

future prices;

business strategies;

customer lists;

cost structures;

algorithms;

source code;

trade secrets.

A CIS must therefore provide:

access controls;

encryption;

audit logs;

role-based permissions;

secure storage;

confidentiality protocols.

28. Cross-Border Competition Intelligence

Modern cartels frequently operate across jurisdictions.

Therefore, competition authorities may need cooperation with:

European Commission;

U.S. DOJ;

FTC;

UK CMA;

CCI;

other national authorities.

International cooperation can facilitate:

information exchange;

simultaneous investigations;

economic analysis;

merger review;

enforcement coordination.

The EU, DOJ and FTC have expressly described cooperation and information exchange as important for technology competition enforcement. (Federal Trade Commission)

29. Important Case Laws

Case 1: Eturas UAB and Others v Lithuanian Competition Council, C-74/14

Facts

Travel agencies used a common online booking system.

The system administrator sent a message concerning an automatic restriction on discounts available through the system.

Issue

Could the use of a common computerised system and the resulting conduct constitute evidence of a concerted practice?

Decision

The Court of Justice addressed the evidentiary requirements for establishing a concerted practice and emphasized that the mere dispatch of the message was not automatically sufficient; the circumstances and evidence had to be assessed, including the presumption of innocence. (Eur-Lex)

Relevance to Competition Intelligence Systems

This is a foundational case for digital competition intelligence.

It demonstrates that:

Digital-system records can be competition evidence, but automated or electronic evidence must still satisfy legal evidentiary requirements.

30. Case 2: United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)

Facts

Microsoft possessed a dominant position in PC operating systems.

The case concerned exclusionary practices involving browsers and competing technologies.

Importance

The court considered:

network effects;

applications barriers;

technological relationships;

distribution arrangements.

Relevance to CIS

A modern intelligence system can monitor precisely these structural factors:

market share;

ecosystem dependence;

interoperability;

distribution;

switching costs.

Microsoft therefore demonstrates why regulators require continuous technological market intelligence, rather than relying only on traditional market-share statistics.

31. Case 3: Google Shopping, Case AT.39740

Facts

The European Commission investigated Google's treatment of comparison-shopping services in its general search results.

Finding

The Commission found that Google had systematically given prominent placement to its own comparison-shopping service while according lower visibility to competing services.

Relevance

The case demonstrates the importance of analysing:

ranking algorithms;

visibility;

traffic;

platform data;

search results.

These are precisely the kinds of variables that competition intelligence systems can monitor.

The case also illustrates the need for regulators to understand algorithmic decision-making rather than merely observing final prices.

32. Case 4: Amazon Marketplace Investigation — UK CMA

Facts

The CMA investigated Amazon's use of third-party seller data, Buy Box selection and Prime-related arrangements.

Outcome

The CMA accepted binding commitments in 2023.

Among other things, Amazon committed not to use rival sellers' Marketplace data to gain an unfair advantage and to ensure equal treatment of offers in Buy Box selection. (GOV.UK)

Relevance

This case demonstrates how competition intelligence must examine:

data flows;

algorithmic rankings;

platform governance;

internal use of competitor information.

It is a strong example of the relationship between data intelligence and platform competition law.

33. Case 5: Bundeskartellamt / Facebook (Meta)

The German competition authority's Facebook proceedings concerned the relationship between Facebook's market position and its processing of user data.

The case demonstrated that competition authorities may need to examine:

data collection;

cross-service data combination;

platform ecosystems;

user dependency;

market power.

Germany's digital competition framework, particularly Section 19a GWB, was designed to permit earlier intervention concerning large digital companies. The Bundeskartellamt has used this framework for companies including Google, Amazon and Facebook. (Bundeskartellamt)

Relevance to CIS

A competition intelligence system can monitor:

users + data + services + market power

rather than examining each service in isolation.

34. Case 6: Intel Corp. v European Commission, C-413/14 P

Facts

The European Commission investigated Intel's conditional rebates.

Importance

The litigation required detailed economic examination of whether the rebates were capable of producing exclusionary effects.

Relevance to CIS

This demonstrates the importance of combining:

pricing data;

customer information;

market shares;

cost information;

economic modelling.

A competition intelligence system can help regulators identify potentially problematic rebate structures, but the ultimate legal analysis requires a proper assessment of competitive effects.

35. Case 7: Airtours plc v Commission, T-342/99

Facts

Airtours proposed acquiring First Choice in the UK package-tour market.

The Commission was concerned about collective dominance.

Decision

The Court annulled the Commission's decision because the evidence did not sufficiently establish the conditions necessary for collective dominance.

Relevance to CIS

This case is an important warning:

Data analytics cannot replace legal and economic proof.

A system may identify:

high concentration;

repeated pricing;

market symmetry.

But these indicators must be connected to legally relevant theories of harm.

36. Case 8: United States v. Apple Inc. — e-books, 952 F. Supp. 2d 638 (S.D.N.Y. 2013)

Facts

The case concerned Apple's role in alleged coordination involving e-book pricing.

Court's analysis

The court examined communications, meetings, agreements and the economic effects of the challenged conduct.

Relevance to CIS

The case illustrates the importance of integrating:

communications evidence;

pricing data;

transaction data;

chronology;

economic evidence.

A modern intelligence system can bring these different categories together to create a coherent investigative timeline.

37. Case 9: Amazon/Deliveroo — UK CMA

Facts

The CMA examined Amazon's minority investment in Deliveroo.

The authority initially identified potential competition concerns, including the possibility that the transaction could affect Amazon's incentives to re-enter or expand in certain markets.

After an in-depth investigation, the CMA cleared the transaction in August 2020. (GOV.UK)

Relevance to CIS

The case illustrates how intelligence systems can support:

merger screening;

potential-competition analysis;

scenario modelling;

market-entry analysis.

It also demonstrates that an initial intelligence signal is not equivalent to a final finding of harm.

38. Competition Intelligence for Cartel Screening

A regulator could create a Cartel Risk Dashboard.

These should be investigative indicators, not automatic infringement findings.

39. Competition Intelligence for Merger Screening

A merger-screening system could calculate:

HHIbeforeHHI_{before} HHIafterHHI_{after}

and:

ΔHHI=HHIafter−HHIbefore\Delta HHI = HHI_{after}-HHI_{before}

It could also identify:

overlapping products;

common customers;

potential competitors;

network effects;

data assets;

patents;

vertical relationships.

40. Competition Intelligence for Dominance

A dominance dashboard might monitor:

Market ShareMarket\ Share Price TrendsPrice\ Trends Entry RatesEntry\ Rates Customer SwitchingCustomer\ Switching ChurnChurn Competitor GrowthCompetitor\ Growth Supplier DependenceSupplier\ Dependence

This can help regulators identify markets requiring deeper examination.

41. AI-Based Competition Intelligence

An advanced system can use machine learning for:

Classification

Categorising complaints.

Anomaly detection

Identifying unusual market behaviour.

Clustering

Grouping similar companies or transactions.

Network analysis

Identifying relationships.

Natural-language processing

Analysing:

contracts;

emails;

complaints;

company statements;

merger documents.

Predictive analytics

Identifying markets likely to require regulatory attention.

However, predictive analytics should be used primarily for prioritisation and investigation support, not automatic legal determinations.

42. Natural-Language Processing

NLP can identify terms associated with potential competition concerns.

For example:

“price increase”

“market allocation”

“exclusive”

“competitor”

“discount”

“customer division”

“bid”

“territory”

It can also analyse large collections of documents more quickly than manual review.

The European Commission maintains extensive structured competition-case information, illustrating the value of searchable case and document systems for enforcement. (Competition Policy)

43. Digital Evidence and Chain of Custody

A competition intelligence system should preserve:

source of data;

date and time;

method of collection;

transformation applied;

analytical method;

analyst intervention;

version history.

This creates a defensible chain of custody.

Without such controls, an automated output may be difficult to rely upon in judicial proceedings.

44. Data Protection and Privacy

Competition intelligence systems can involve personal information.

For example:

employee communications;

consumer complaints;

user data;

location information.

Therefore, regulators must balance:

Competition enforcement

with:

privacy and data-protection obligations.

Data collection should be:

lawful;

necessary;

proportionate;

secure;

purpose-limited.

45. Cybersecurity

Competition authorities possess highly sensitive information.

A successful cyberattack could expose:

merger plans;

trade secrets;

investigation strategies;

confidential company information.

Therefore, CIS infrastructure requires:

encryption;

identity management;

access controls;

logging;

intrusion detection;

secure backups.

46. International Intelligence Sharing

Competition problems increasingly cross borders.

For example:

Company A — India

Company B — EU

Company C — USA

Customers — worldwide

A coordinated intelligence framework can help authorities identify common conduct.

International cooperation can involve:

compatible analytical methods;

secure information exchange;

simultaneous investigations;

economic evidence;

merger analysis.

EU and U.S. authorities have expressly identified technology-sector cooperation and information exchange as important to addressing modern competition challenges. (Federal Trade Commission)

47. Competition Intelligence and CCI

For the Competition Commission of India, a CIS can support:

Section 3

Detection of anti-competitive agreements.

Section 4

Monitoring potential abuse of dominant position.

Sections 5–6

Combination and merger screening.

Section 19

Market inquiries and investigations.

Section 26

Investigation support.

Section 27

Evidence-supported enforcement decisions.

Section 49

Competition advocacy and market awareness.

The CCI maintains a dedicated database of judgments and legal materials, providing a foundation for structured legal intelligence. (Competition Commission of India)

48. Benefits of Competition Intelligence Systems

1. Faster detection

Large markets can be monitored continuously.

2. Better resource allocation

Authorities can prioritise high-risk markets.

3. Early intervention

Potential problems can be identified before substantial harm occurs.

4. Evidence integration

Different data sources can be analysed together.

5. Digital-market monitoring

Complex platforms can be monitored more effectively.

6. Cartel detection

Suspicious patterns can be identified.

7. Merger screening

Transactions can be prioritised.

8. Remedy monitoring

Compliance can be monitored continuously.

49. Risks and Limitations

A. False positives

A lawful pattern may appear suspicious.

B. False negatives

Sophisticated violations may evade detection.

C. Algorithmic bias

Historical enforcement data may contain biases.

D. Lack of explainability

Black-box systems can undermine procedural fairness.

E. Privacy risks

Large-scale data collection can create privacy problems.

F. Over-reliance on technology

Competition law remains a legal and economic discipline.

G. Data-quality problems

Incomplete or incorrect data can produce misleading conclusions.

50. Best-Practice Regulatory Model

An effective system should follow:

Stage 1 — Collection

Gather lawful market information.

↓

Stage 2 — Cleaning

Remove duplicate and erroneous data.

↓

Stage 3 — Integration

Connect multiple datasets.

↓

Stage 4 — Analytics

Apply economic and computational methods.

↓

Stage 5 — Risk scoring

Identify potential concerns.

↓

Stage 6 — Human review

Economists and lawyers examine alerts.

↓

Stage 7 — Investigation

Use statutory investigative powers.

↓

Stage 8 — Evidence verification

Test the underlying facts.

↓

Stage 9 — Legal assessment

Apply the relevant competition-law provisions.

↓

Stage 10 — Decision

Issue the appropriate order or close the matter.

↓

Stage 11 — Monitoring

Monitor compliance and market developments.

51. Competition Intelligence and Regulatory Governance

A competition intelligence system should itself be governed by principles of:

accountability;

transparency;

proportionality;

independence;

security;

explainability;

auditability;

human oversight.

This is particularly important because regulators exercise coercive legal powers.

An automated risk score should therefore assist a regulator, not become the regulator.

52. Competition Intelligence and AI

The issue is becoming particularly important with AI.

AI systems can:

analyse millions of documents;

identify pricing patterns;

detect unusual bidding;

map corporate relationships;

monitor digital platforms;

examine contracts.

At the same time, AI itself can create competition problems through:

algorithmic pricing;

coordination;

exclusion;

discriminatory ranking;

concentration of computing resources;

concentration of data.

The FTC, DOJ and international competition authorities have explicitly recognised competition risks associated with AI and the need for enforcement across the AI ecosystem. (Federal Trade Commission)

53. Difference Between Competition Intelligence and Surveillance

This distinction is important.

Competition intelligence

Legally authorised, purpose-specific collection and analysis of information for competition enforcement.

General surveillance

Broad monitoring without sufficiently defined competition-law purposes or safeguards.

A legitimate CIS should therefore have:

statutory authority;

defined purposes;

proportionality;

confidentiality;

judicial/administrative safeguards where required.

54. Key Principles Emerging from the Case Law

The cases discussed establish several important principles.

Eturas

Electronic system records can constitute competition evidence, but evidentiary standards and the presumption of innocence remain important. (Eur-Lex)

Microsoft

Technology markets require analysis of network effects, interoperability and barriers to entry.

Google Shopping

Competition analysis may require examination of ranking and algorithmic visibility.

Amazon Marketplace

Data use and algorithmic platform decisions can become competition issues. (GOV.UK)

Intel

Complex pricing practices require detailed economic analysis rather than simplistic indicators.

Airtours

High concentration or analytical indicators do not automatically establish coordinated conduct.

Apple e-books

Economic evidence must be integrated with communications and documentary evidence.

Amazon/Deliveroo

Intelligence signals and preliminary concerns must be distinguished from final findings after detailed investigation. (GOV.UK)

55. Short Examination Answer

Competition Intelligence Systems for Regulators are technological and analytical frameworks used by competition authorities to collect, integrate and analyse market information for detecting possible anti-competitive conduct. They can monitor prices, bids, market shares, mergers, corporate relationships, consumer complaints, digital-platform behaviour, algorithms and data concentration.

Their major functions include cartel detection, bid-rigging detection, abuse-of-dominance monitoring, merger screening, digital-market surveillance, market studies and remedy monitoring. Artificial intelligence, machine learning, natural-language processing, network analysis and anomaly detection can improve regulatory capacity.

Important cases include Eturas, Microsoft, Google Shopping, Amazon Marketplace, Intel, Airtours, Apple e-books and Amazon/Deliveroo. Eturas is particularly significant because it shows that evidence generated through a common digital booking system can be relevant to establishing a concerted practice, while also emphasising evidentiary safeguards. (Eur-Lex)

However, an algorithmic alert is not equivalent to proof of a competition-law infringement. Human review, lawful evidence gathering, procedural fairness, explainability, confidentiality and proportionality remain essential.

56. Conclusion

Competition Intelligence Systems represent the transition from reactive competition enforcement to data-supported and potentially proactive market monitoring.

The basic model is:

Market data → analytical intelligence → risk identification → human investigation → legal evidence → enforcement → continuous monitoring

Their importance is greatest in modern markets characterised by:

algorithms;

digital platforms;

network effects;

large datasets;

rapid price changes;

complex corporate structures;

cross-border transactions.

Cases such as Eturas, Microsoft, Google Shopping, Amazon Marketplace, Intel, Airtours and Apple e-books demonstrate why regulators increasingly need to understand both technology and economics when enforcing competition law.

At the same time, competition intelligence must remain subordinate to legal standards. Data patterns can generate leads, but they should not replace proof; AI can assist investigators, but it should not make unreviewable legal conclusions. The most appropriate regulatory model is therefore a human-supervised, explainable, secure and legally accountable intelligence system that strengthens enforcement while protecting due process and confidential business information.