Competition Law and Machine-Managed Allocation Mechanisms

Detailed Explanation with At Least 6 Case Laws — Indian Competition Law

1. Introduction

Machine-managed allocation mechanisms are systems in which algorithms, artificial intelligence, machine-learning models, or automated software determine how scarce commercial opportunities are allocated among suppliers, distributors, customers, drivers, sellers, advertisers, service providers, or other market participants.

Examples include algorithms deciding:

which seller receives a customer;

which supplier gets a particular territory;

which driver receives a ride request;

which advertiser receives an advertising opportunity;

which seller receives platform visibility;

which distributor receives inventory;

which bidder receives a contract;

which customer receives a particular offer;

which competitor receives or is denied access to a platform.

The technology itself is not unlawful. The competition issue arises when the allocation mechanism restricts competition, facilitates coordination, discriminates in favour of a dominant enterprise, forecloses rivals, or allocates markets among competitors.

The Competition Commission of India (CCI) has expressly recognised AI as a technology capable of changing competitive dynamics and released its Market Study on Artificial Intelligence and Competition in October 2025. (Competition Commission of India)

2. Meaning of Machine-Managed Allocation

A useful definition is:

Machine-managed allocation means the automated distribution of customers, resources, transactions, territories, supply, visibility, opportunities, or other commercially valuable inputs among market participants according to rules or predictions generated by software, algorithms, or AI.

Simple example

Suppose an online marketplace has 10,000 sellers.

Its AI system automatically decides:

Customer searches for Product X → Algorithm selects Seller A.

If the algorithm uses legitimate factors such as:

price;

delivery time;

product quality;

stock availability;

there may be significant efficiency benefits.

But if the algorithm is deliberately programmed to:

“Never allocate customers to sellers who use competing platforms,”

the competition implications become much more serious.

3. Allocation Is Not Automatically Anti-Competitive

The first principle is important:

Machine-managed allocation ≠ automatic competition-law violation.

Businesses constantly allocate:

inventory;

customers;

advertising;

distribution territories;

delivery capacity;

production capacity.

Competition law becomes relevant when the allocation mechanism has an anti-competitive purpose or effect falling within the statutory framework.

Under the Competition Act, 2002, Section 3 deals with anti-competitive agreements, Section 4 with abuse of dominant position, and the Act also regulates combinations. (Competition Commission of India)

4. Section 3 and Machine-Managed Allocation

Section 3 prohibits agreements that cause or are likely to cause an appreciable adverse effect on competition (AAEC).

This is particularly important when the machine is used to implement an agreement between businesses.

For example:

Manufacturer A + Manufacturer B + common algorithm → customers divided between A and B.

The software may merely be the mechanism through which the agreement is implemented.

The legal analysis therefore focuses on the underlying arrangement rather than giving importance only to the technological instrument.

5. Section 3(3): Market Allocation

Section 3(3) is particularly relevant because agreements among competitors to allocate markets or customers are among the categories presumed to have AAEC, subject to the statutory framework.

CCI identifies market allocation as one of the four principal horizontal restrictions under Section 3(3), along with price fixing, limiting supply/production, and bid rigging. (Competition Commission of India)

Example

Three competing logistics companies use a common algorithm:

Company A → Delhi

Company B → Mumbai

Company C → Bengaluru

If the allocation reflects coordination among competitors rather than independent commercial decisions, competition concerns arise.

6. Section 3(4): Vertical Allocation

Machine-managed allocation can also arise between enterprises at different levels of the supply chain.

Section 3(4) covers vertical restraints such as:

exclusive supply;

exclusive distribution;

refusal to deal;

tie-in arrangements;

resale price maintenance.

CCI expressly identifies these as vertical restraints. (Competition Commission of India)

For example, a manufacturer might use an automated system to allocate particular territories exclusively to selected distributors.

7. Section 4: Allocation by a Dominant Enterprise

Section 4 becomes particularly important where the allocation mechanism is operated by a dominant enterprise.

Dominance itself is not prohibited.

The concern arises where dominance is abused.

Relevant forms of abuse may include:

denying market access;

limiting markets;

limiting technical development;

imposing unfair conditions;

imposing unrelated contractual conditions;

leveraging dominance from one market into another.

CCI identifies these categories in its explanation of Section 4. (Competition Commission of India)

8. Algorithmic Allocation and Market Access

Consider a dominant online marketplace.

Its algorithm allocates 95% of customer traffic to its own affiliated sellers and only 5% to independent sellers.

The allocation mechanism could potentially affect:

seller access;

consumer choice;

competitor viability;

entry;

innovation.

The question would be whether the algorithm is producing legitimate efficiency-based results or is being used to exclude competitors or distort competition.

9. Allocation of Customers

Customer allocation is especially sensitive.

An algorithm may determine which customer receives:

a product;

a service provider;

a driver;

a bank offer;

an insurance product;

an advertisement;

a loan;

a telecommunications service.

Where competing enterprises independently use algorithms to allocate customers, there may be no competition issue.

But where competitors coordinate customer allocation through a common system, the matter can potentially become a horizontal market-allocation issue.

10. Algorithmic Territory Allocation

A machine can automatically divide geographic markets.

For example:

If these are independent commercial decisions based on logistics, the arrangement may be legitimate.

If competing enterprises have agreed to divide markets and the algorithm implements that agreement, Section 3(3) concerns may arise.

11. Algorithmic Customer Allocation

The same principle applies to customers.

An algorithm might allocate:

large customers to Firm A;

medium customers to Firm B;

small customers to Firm C.

If competitors have agreed to this allocation, it could potentially amount to customer-market allocation.

The fact that the agreement is implemented through software does not necessarily change its legal character.

12. Machine-Managed Allocation and Bid Rigging

Automated allocation systems can also affect procurement.

Imagine five contractors using a common software platform.

The system determines:

Contractor A wins Tender 1.
Contractor B wins Tender 2.
Contractor C wins Tender 3.

If this results from independent optimisation, it may be harmless.

But if competing bidders coordinate through the system to determine who will win which tender, the arrangement may raise bid-rigging/collusive bidding concerns.

13. Machine Allocation and Information Exchange

Algorithms frequently require large quantities of information.

A system might process:

prices;

capacity;

inventory;

customer demand;

future strategies;

bids;

production levels.

If competitors share competitively sensitive information through a common platform, competition risks can increase.

The crucial question is whether the information exchange facilitates coordination or otherwise reduces strategic uncertainty between competitors.

14. Common Algorithm as a Coordination Mechanism

A particularly important scenario is:

Competitor A → Common algorithm ← Competitor B

Both competitors may independently enter information into the system.

If the algorithm then recommends similar conduct to both, authorities may examine whether the system facilitates coordinated behaviour.

This is sometimes described as a hub-and-spoke type risk.

However, merely using the same software is not automatically proof of an illegal agreement. Evidence concerning communication, knowledge, adoption and concerted conduct remains important.

15. Machine-Managed Allocation in Digital Platforms

Digital platforms are particularly suitable for algorithmic allocation.

A platform may determine:

which seller appears first;

which driver gets a ride;

which restaurant receives an order;

which advertisement is displayed;

which app receives visibility;

which product receives recommendation.

The platform therefore controls an important allocation layer between buyers and sellers.

That allocation power can become commercially significant when the platform is difficult to bypass.

16. Network Effects

Machine-managed allocation can reinforce network effects.

For example:

More sellers → more products → more consumers → more transactions → more data → better algorithm → more consumers → more sellers.

If the algorithm then systematically directs transactions toward the platform's own businesses, competitors may have difficulty achieving sufficient scale.

Thus, allocation can become an important mechanism for ecosystem entrenchment.

17. Data Advantage

Algorithms improve as they receive more data.

A dominant platform may possess:

transaction data;

consumer preferences;

seller performance data;

conversion rates;

pricing information;

demand forecasts.

The platform can use this information to improve its allocation decisions.

This can create a feedback loop:

More transactions → more data → better algorithm → better allocation → more transactions.

Such data-driven advantages are among the competition issues considered in the CCI's work on AI and competition. (Competition Commission of India)

18. Self-Preferential Allocation

One major concern is self-preferencing.

Suppose a platform owns its own retail business.

Its algorithm determines:

Platform-owned product → priority allocation
Independent seller → ordinary allocation

The platform may therefore use its control over the allocation mechanism to favour its own downstream business.

The legal assessment may involve:

dominance;

denial of market access;

leveraging;

preferential treatment;

vertical foreclosure.

19. Google Android and Ecosystem Allocation

Case: Umar Javeed & Others v. Google LLC & Another

CCI Case No. 39/2018

The CCI proceedings concerning Google's Android ecosystem examined arrangements involving Google's proprietary services and device manufacturers. The CCI opened the matter under the Competition Act's antitrust framework and later issued its detailed Android decision. (Competition Commission of India)

Relevance

Although this was not a case specifically about an AI allocation algorithm, it demonstrates how CCI can examine technological ecosystems and contractual arrangements that influence access and distribution.

For machine-managed allocation, the analogous issue is:

Does the technological system determine access to an important ecosystem in a way that disadvantages competing services?

20. Delhi Vyapar Mahasangh v. Flipkart

Case: Delhi Vyapar Mahasangh v. Flipkart Internet Pvt. Ltd. & Ors.

CCI Case No. 40/2019

CCI's investigation order concerned allegations involving, among other things:

exclusive launches;

preferred sellers;

deep discounting;

preferential listing/promotion.

The CCI directed investigation under Section 26(1) concerning alleged vertical restraints. (Competition Commission of India)

Relevance to machine-managed allocation

An algorithm could automatically determine:

which sellers receive preferred listing;

which sellers receive discounts;

which products receive visibility;

which sellers receive customer traffic.

Thus, the case provides a useful framework for understanding how automated allocation of marketplace opportunities can affect competition.

21. Preferred Seller Allocation

The Flipkart proceedings are especially relevant to algorithmic allocation because preferential treatment can be implemented technologically.

Suppose an algorithm identifies a group of "preferred sellers" and automatically gives them:

higher rankings;

more customer traffic;

lower commissions;

better inventory access.

The competition issue is whether this allocation merely reflects legitimate efficiency or materially disadvantages competing sellers.

The CCI's investigation order recorded allegations concerning preferential sellers and preferential listing. (Competition Commission of India)

22. All India Online Vendors Association v. Flipkart

Case: All India Online Vendors Association v. Flipkart India Pvt. Ltd. & Ors.

CCI Case No. 20/2018

This proceeding is relevant to marketplace competition and the relationship between platforms and sellers.

Importance

It helps demonstrate why control over a marketplace's allocation mechanisms can have competitive significance.

Where a platform determines which sellers receive meaningful access to consumers, the allocation mechanism can become an important competitive parameter.

23. Amazon Private-Label Proceedings

Case: In Re: Allegations pertaining to private label brands related to Amazon sold on Amazon India marketplace

Suo Motu Case No. 04/2021

The proceedings concerned allegations surrounding Amazon's marketplace and private-label arrangements.

Relevance

A marketplace that simultaneously:

controls the allocation mechanism,

operates competing products,

possesses extensive seller data,

may raise competition questions concerning whether its allocation system is neutral.

A machine-learning system can make these issues more complex because allocation decisions may be embedded within ranking, recommendation and advertising systems.

24. Matrimony.com v. Google

Case: Matrimony.com Ltd. v. Google LLC & Ors.

CCI Case Nos. 07/2012 and 30/2012

The proceedings concerned Google's search-related conduct and allegations concerning preferential treatment.

Relevance

The broader lesson for machine-managed allocation is that an algorithmic ranking or allocation mechanism can influence competitive opportunities.

Where a powerful intermediary controls access to consumers, algorithmic placement can itself become commercially important.

25. CCI v. SAIL

Case: Competition Commission of India v. Steel Authority of India Ltd.

(2010) 10 SCC 744

This is a foundational Supreme Court decision on CCI's investigation process.

Relevance

Machine-managed allocation cases can be highly technical.

CCI may need to examine:

software;

algorithms;

internal policies;

contracts;

databases;

system logs;

communications;

decision rules.

The SAIL framework remains important for understanding the threshold at which CCI may proceed with an investigation.

26. Excel Crop Care Ltd. v. CCI

Case: Excel Crop Care Ltd. v. Competition Commission of India

(2017) 8 SCC 47

The Supreme Court dealt with competition-law infringement and penalty principles.

Relevance

The case demonstrates that competition law evaluates the economic and legal substance of conduct, rather than allowing enterprises to escape scrutiny through the particular form in which conduct is implemented.

This principle is relevant where an enterprise says:

“The algorithm made the decision, not the company.”

The technological mechanism does not by itself determine legality.

27. CCI v. Bharti Airtel

Case: Competition Commission of India v. Bharti Airtel Ltd.

(2019) 2 SCC 521

The Supreme Court examined the relationship between sector-specific regulation and competition law.

Relevance

Machine-managed allocation is increasingly used in regulated industries such as:

telecommunications;

banking;

insurance;

payments;

transportation.

Therefore, competition authorities may need to consider both the Competition Act and the regulatory framework governing the relevant sector.

28. Eturas v. Lithuanian Competition Authority

Case: Eturas UAB and Others v. Lietuvos Respublikos konkurencijos taryba

Case C-74/14

The European Court of Justice considered competition issues involving a common computerised booking system.

Importance

The case is highly relevant conceptually because a computerised system can become part of the mechanism through which coordinated commercial conduct occurs.

For machine-managed allocation, this raises questions concerning:

common software;

automated instructions;

system messages;

user knowledge;

adoption of algorithmic recommendations;

coordinated conduct.

It is an international analogical authority rather than an Indian precedent.

29. United States v. Microsoft

Case: United States v. Microsoft Corp.

253 F.3d 34 (D.C. Cir. 2001)

The case concerned Microsoft's conduct in the technological ecosystem surrounding operating systems and browsers.

Relevance

It illustrates how technological control over an important platform can affect competitive opportunities in adjacent markets.

For machine-managed allocation, the analogous question is:

Can control over the technological allocation layer be used to disadvantage competing businesses?

30. Machine-Managed Allocation and Refusal to Deal

An allocation system can effectively become a refusal-to-deal mechanism.

For example:

Algorithm automatically rejects every supplier that also supplies a competing platform.

If the enterprise has significant market power, such conduct could potentially raise Section 4 issues.

The analysis would depend upon:

dominance;

market definition;

necessity of access;

foreclosure;

legitimate business justification;

competitive effects.

31. Allocation of Scarce Inventory

Suppose there are only 1,000 units of a critical product.

An algorithm decides which distributors receive inventory.

If the algorithm is based on:

demand;

delivery capacity;

creditworthiness;

historical performance,

the allocation may be commercially rational.

But if the system systematically denies inventory to distributors dealing with competitors, the allocation can become exclusionary.

32. Allocation of Advertising Opportunities

Digital advertising provides another example.

An advertising algorithm determines:

Which advertiser receives an impression?

Where the platform controls both:

the advertising marketplace, and

competing advertising inventory,

algorithmic allocation can raise questions regarding:

self-preferencing;

discriminatory access;

tying;

exclusion;

data advantages.

33. Allocation of Drivers and Rides

Ride-hailing platforms use algorithms to allocate drivers to passengers.

The system may consider:

distance;

driver availability;

traffic;

passenger demand;

driver rating.

These are normally efficiency considerations.

Competition concerns could arise if competing ride platforms coordinate through a common system to allocate drivers or customers between themselves.

34. Allocation in Labour Platforms

Algorithms may allocate work to gig workers.

Competition issues could arise where:

multiple competing platforms use common algorithms;

worker availability is restricted;

workers are prevented from using rival platforms;

sensitive compensation information is exchanged;

platforms coordinate allocation.

This creates an important intersection between competition law and labour-market competition.

35. Allocation and Algorithmic Discrimination

Machine-managed allocation may discriminate between market participants based on:

price;

location;

historical behaviour;

platform affiliation;

customer profile;

competitor relationships.

Not every difference is legally discriminatory in the competition-law sense.

The competition question is whether the differentiated treatment produces an exclusionary or exploitative effect protected or prohibited by the Competition Act.

36. Dynamic Allocation

Unlike traditional contractual arrangements, machine allocation can change continuously.

For example:

9:00 AM: Seller A receives 70% of traffic.
12:00 PM: Seller B receives 60%.
3:00 PM: Seller C receives most traffic.

The algorithm may respond to real-time data.

This creates a challenge for competition analysis because the relevant conduct may not be one fixed decision.

Authorities may need to examine the pattern of decisions over time.

37. Algorithmic Allocation and Predatory Conduct

A dominant enterprise could theoretically use algorithmic allocation to:

identify emerging competitors;

reduce their access to customers;

increase the dominant firm's own allocation;

use temporary discounts to retain customers;

restore higher prices after competitors weaken.

Such conduct would require careful economic analysis.

The existence of aggressive competition alone does not establish predatory conduct.

38. Allocation and Entry Barriers

A new entrant often needs access to:

customers;

suppliers;

distributors;

data;

infrastructure;

payment systems;

advertising.

If a dominant algorithm controls all these access points, new entrants may face substantial barriers.

Thus:

Control over allocation can become control over market entry.

39. Machine Allocation and Innovation

Competition is not limited to current prices.

An allocation system may affect:

innovation;

product quality;

business models;

new entrants;

technological development.

If an algorithm systematically allocates customers away from innovative competitors, the immediate price impact may be small while the long-term innovation effect is significant.

40. Efficiency Defence

Businesses may argue that machine allocation creates efficiencies.

Examples:

faster matching;

reduced transaction costs;

better inventory utilisation;

lower logistics costs;

improved customer experience;

reduced wastage;

better capacity utilisation;

improved product availability.

These benefits should be considered when assessing whether conduct actually produces AAEC.

41. Relevant Factors for CCI Analysis

A machine-managed allocation mechanism can be examined through:

1. Relevant market

Where does the competition occur?

2. Market power

Does the enterprise have significant market power?

3. Allocation criteria

What factors does the algorithm use?

4. Competitor impact

Which competitors are affected?

5. Market coverage

How much of the market is controlled by the system?

6. Duration

How long does the effect continue?

7. Switching costs

Can businesses move elsewhere?

8. Network effects

Does allocation reinforce an existing ecosystem?

9. Data advantage

Does the operator possess superior information?

10. Efficiency

Are there legitimate business justifications?

42. Allocation Mechanism as an Essential Bottleneck

In some digital markets, the allocation system can become a bottleneck.

For example:

Consumers → Platform Algorithm → Sellers

If consumers cannot practically bypass the platform, sellers depend on the platform's allocation decisions.

The allocation algorithm therefore becomes commercially critical.

Where the platform is dominant, discriminatory access may raise Section 4 concerns.

43. Transparency

One important issue is whether businesses understand why they receive particular allocations.

A seller might receive very little traffic without knowing:

which algorithmic factor caused it;

whether competitors received preferential treatment;

whether platform affiliation affected ranking;

whether the system imposed hidden restrictions.

Transparency is therefore increasingly important for competition compliance.

44. Human Oversight

Businesses operating important allocation systems should maintain human oversight.

A sensible compliance structure could include:

Algorithm design → Competition review → Testing → Deployment → Monitoring → Audit → Corrective action

This helps identify whether the system is unintentionally producing exclusionary outcomes.

45. Algorithmic Audit Trail

A company should preserve:

algorithm versions;

model changes;

decision logs;

allocation rules;

training data categories;

contractual conditions;

human approvals;

exception decisions;

complaints and responses.

This can be particularly important during a CCI investigation.

46. Machine-Managed Allocation and Mergers

Allocation power can also become relevant in merger control.

Suppose two major platforms merge and the combined entity controls:

consumer data;

seller access;

advertising;

logistics;

allocation algorithms.

The CCI's combination analysis includes factors such as barriers to entry, market concentration, countervailing power, vertical integration, innovation and the likelihood of removing an effective competitor. (Competition Commission of India)

Thus, acquisition of an important allocation platform can potentially create competition concerns even before an exclusionary practice occurs.

47. Remedies

If an allocation mechanism produces competition concerns, possible remedies may include:

Structural remedies

divestiture;

separation of businesses.

Behavioural remedies

prohibit discriminatory allocation;

require equal access;

prohibit exclusionary conditions;

modify ranking criteria.

Technical remedies

algorithmic audits;

independent monitoring;

access logs;

model documentation.

Transparency remedies

explain allocation criteria;

disclose important commercial conditions;

provide appeal mechanisms.

The appropriate remedy depends on the facts and statutory framework.

48. Traditional Allocation vs Machine-Managed Allocation

49. Major Competition Risks

Machine-managed allocation may create:

Market allocation

Customer allocation

Territorial allocation

Bid allocation

Supplier allocation

Inventory allocation

Preferential allocation

Self-preferential allocation

Exclusionary allocation

Algorithmic coordination

Information-exchange risks

Entry barriers

Network-effect reinforcement

Data-based competitive advantages

Platform foreclosure

50. Important Case-Law Summary

51. Key Legal Principles

Automated allocation is not inherently anti-competitive.

The Competition Act applies regardless of whether a decision is made by a person or an algorithm.

Section 3 is important where allocation implements an anti-competitive agreement.

Section 3(3) is particularly relevant to market/customer allocation between competitors.

Section 3(4) can apply to vertical allocation arrangements.

Section 4 is important where a dominant enterprise uses allocation to exclude competitors.

Control over allocation can become control over market access.

Data can make allocation systems more powerful.

Network effects can amplify allocation advantages.

Common algorithms can create coordination risks.

The same software being used by competitors does not, by itself, prove collusion.

Efficiency benefits must be considered.

Algorithmic allocation should be assessed over time where decisions are dynamic.

Platform self-preferencing can make allocation systems particularly significant.

Algorithmic transparency and auditability are increasingly important.

Merger control can address accumulation of allocation power before actual exclusion occurs.

52. Quick Revision

Meaning

Machine-managed allocation = automated distribution of customers, resources, transactions, supply, territory or opportunities through algorithms/AI.

Main legal provisions

Section 3(1) — prohibition of agreements causing or likely to cause AAEC

Section 3(3) — horizontal market allocation and related restrictions

Section 3(4) — vertical restraints

Section 4 — abuse of dominant position

Section 19 — inquiry by CCI

Section 26 — investigation process

Section 20 — inquiry into combinations

Basic competition chain

Algorithm → Allocation → Market Access → Foreclosure/Coordination → Competitive Effect

Most important cases

CCI v. SAIL

Excel Crop Care v. CCI

CCI v. Bharti Airtel

Umar Javeed v. Google

Delhi Vyapar Mahasangh v. Flipkart

AIOVA v. Flipkart

Matrimony.com v. Google

Amazon Private-Label proceedings

Eturas

Microsoft

Conclusion

Machine-managed allocation mechanisms are becoming an important competition-law issue because algorithms increasingly control who receives customers, products, distribution opportunities, advertising exposure, inventory and access to digital ecosystems.

The central legal principle is that technology does not change the fundamental competition-law question. The CCI must examine whether the allocation is an independent and efficiency-enhancing business decision or whether it implements or facilitates conduct that restricts competition.

The greatest risks arise where a powerful platform or group of competitors uses an algorithm to allocate markets, coordinate conduct, discriminate against rivals, restrict market access, favour affiliated businesses, or reinforce network and data advantages. India's existing Sections 3 and 4 provide the principal framework, while the CCI's 2025 AI and Competition Market Study demonstrates that the Commission is actively examining how AI changes competitive conditions. (Competition Commission of India)

In short:

The machine may manage the allocation, but competition law examines the competitive consequences of the allocation.