Section 3(k) of the Patents Act: Where Software Patents Stand in India

Section 3(k) of the Patents Act, 1970 is the single most-litigated subject-matter exclusion in Indian patent practice. The provision states that 'a mathematical or business method or a computer programme per se or algorithms' are not inventions within the meaning of the Act. The words 'per se' carry the entire weight of Indian software-patent jurisprudence. Their interpretation — through the Computer-Related Inventions (CRI) Guidelines of 2017 and the Delhi High Court's Ferid Allani v. Union of India decision in 2019 — decides what kind of software innovation can be patented in India and what cannot.

This guide covers the text of Section 3(k), the role of the CRI Guidelines, the Ferid Allani holding, and how Indian examiners actually apply the section in 2026.

The text and the 'per se' problem

Section 3(k) excludes four categories: mathematical methods, business methods, computer programmes per se, and algorithms. 'Per se' applies only to computer programmes — not to the other three. The other three are excluded outright. A computer programme by itself is excluded; a computer programme that, in combination with hardware or as part of a larger system, produces a technical effect that goes beyond the routine running of a program may be patentable.

The drafting history of the 'per se' qualifier matters. A 2002 Joint Parliamentary Committee inserted the qualifier specifically to ensure that incidental use of a computer programme in an otherwise patentable invention did not exclude the whole invention. The intent was narrowing, not abolishing, software-related claim space.

A computer programme is excluded. A computer programme that produces a technical effect is not.

Ferid Allani v. Union of India (2019)

The Delhi High Court in Ferid Allani v. Union of India, W.P.(C) 7/2014, decided in December 2019, reviewed the Patent Office's refusal of an application titled 'Method and Device for Accessing Information Sources and Services on the Web'. The Court remanded the application for re-examination and laid down clear interpretive principles:

• 'Per se' must be read with purpose. A computer programme with no technical effect is excluded; one that produces a technical effect beyond the normal running of a programme is not
• The test is whether the invention produces a 'technical contribution' or 'technical effect' — a phrase imported into Indian practice from European jurisprudence
• Almost every modern invention uses software in some form. A reading that excluded any software-implemented invention would defeat the patent system's purpose

Ferid Allani is now the leading Indian authority on Section 3(k). Indian examiners are required to engage with the 'technical effect' analysis rather than reject software-related claims summarily.

The 2017 CRI Guidelines

The Patent Office's Computer-Related Inventions Examination Guidelines of 2017 set out a three-step test for examiners:

1. Identify the actual contribution claimed by the invention
2. If the contribution lies in a mathematical method, business method or algorithm — refuse under Section 3(k)
3. If the contribution lies in a computer programme per se — refuse. If the contribution lies in a method or article that uses a computer programme but achieves a technical effect outside the programme itself — proceed to examine novelty, inventive step and industrial applicability

The 2017 Guidelines replaced earlier 2016 Guidelines that had included a 'novel hardware' requirement, which was withdrawn after industry pushback. The current Guidelines are the operational manual every Indian software-related patent application is examined under.

What India will grant in 2026

Indian software-related patents are granted in well-defined categories:

• Network and communication inventions — routing improvements, encoding/decoding methods, signal processing
• Security inventions — encryption schemes, authentication protocols, intrusion detection methods producing measurable security gain
• Embedded systems — control systems for industrial machinery, automotive electronics, IoT devices
• Image and signal processing — compression, recognition, enhancement methods with measurable technical improvement
• Database and storage architecture — methods that reduce memory usage, improve retrieval times, optimise indexing

What India will not grant: pure business methods (revenue models, pricing schemes, financial-transaction logic), abstract algorithms without technical effect, mathematical methods, and claims that read as a user interface or UX flow without technical underpinning.

The drafting playbook

Three drafting moves that change the outcome of a Section 3(k) examination:

1. Anchor every independent claim to a technical effect. The first phrase of the claim should describe what the invention does technically — not what software runs inside it
2. Disclose the technical problem and the technical solution in the specification. The examiner needs to see in the description, not infer from the claims, that a technical contribution exists
3. Tie software claims to hardware where possible. System claims that recite 'a processor configured to' or 'a memory storing instructions that when executed' anchor the software to a tangible apparatus, even though the 2016 'novel hardware' requirement was dropped

Comparison with US and EU practice

Indian Section 3(k) sits between the European 'technical effect' doctrine and the US Alice/Mayo two-step test. The European Patent Office's Article 52(2) and (3) exclusion of computer programmes 'as such' is the closest analogue — and Indian courts have explicitly drawn on EPO jurisprudence in interpreting 'per se'. US claims that have survived Alice typically also survive Indian 3(k); EPO claims that survive the technical-effect test do as well. Pure-software claim language that survives in the US can still fail Section 3(k) in India — the Indian bar is closer to the European than the American.

The takeaway

Section 3(k) is a real exclusion, but a narrow one after Ferid Allani. Software-related inventions with a clearly articulated technical effect — security, performance, communication, control, signal-processing — continue to be granted Indian patents in 2026. Pure business-method and algorithm claims continue to fail. The decision is in the drafting, not the underlying technology. IPForte's complete patent filing service handles CRI-track applications regularly; office-action responses on Section 3(k) are one of our most common matters.