Software and AI Patents in India
You have a software product, an AI model, or a computer-implemented invention you want to patent in India. The decisive question is not whether code is involved; it is whether the claimed subject matter solves a technical problem in a patentable way, anchored in a demonstrable technical effect, rather than reading as a computer programme, algorithm, mathematical method, or business method per se. Intepat's Registered Patent Agents handle Section 3(k) analysis, specification drafting, FER responses, and hearing-stage prosecution before the Indian Patent Office.
Speak to a Patent AgentPatent Filing Timeline: India
What software and AI patent filing covers
Intepat handles the full prosecution cycle for software, AI, and computer-implemented inventions:
Patentability assessment to determine whether the invention can be framed as a technical solution rather than excluded subject matter
Specification drafting around a demonstrable technical effect, with claim sets structured for Indian examination practice
Provisional and complete specification filing
FER response work addressing Section 3(k) objections through technical-effect arguments, claim amendments, and prior-art analysis
Section 14 hearing representation through to grant
PCT preparation where international protection is needed
Coordinated claim-drafting across Indian, USPTO, and EPO jurisdictions through foreign associates
Section 3(k) and the CRI Guidelines
Section 3(k) of the Patents Act, 1970
Section 3(k) of the Patents Act, 1970 excludes a mathematical method, a business method, a computer programme per se, and algorithms from patentability. The operative phrase is "per se": an invention is not excluded simply because it involves software or AI. The exclusion applies when the claim covers the programme, algorithm, or method in isolation, without technical character anchored in a real-world outcome. Claims drafted around a business outcome, a mathematical optimisation without technical grounding, or an algorithm that operates identically on any generic processor are vulnerable to rejection. Claims anchored in a specific technical problem and its technical solution are assessed on novelty and inventive step.
CRI Guidelines and technical-effect doctrine
The Guidelines for Examination of Computer Related Inventions direct examiners to assess the substance of what the invention contributes. Technical effects recognised in examination practice include improvements in processing efficiency, reductions in memory usage, improvements in network communication reliability, and novel physical apparatus controlled by the algorithm.
AI and machine-learning inventions require particular attention. A claim describing a neural-network architecture without identifying the specific technical problem it addresses and how the architecture contributes to a measurable outcome typically attracts a Section 3(k) objection. Technical effects that have supported AI prosecution include improved accuracy in sensor signal processing, reduced false positives in a constrained technical environment, resource-efficient model execution under defined hardware constraints, and network security anomaly detection with measurable performance improvement.
Claims are typically framed as method claims tied to a specific technical process or as apparatus claims around a configured system, rather than as product claims around the programme or model itself.
Stronger and weaker patent candidates
Stronger candidates are claims directed to a specific technical outcome: an AI-enabled physical sensing system, embedded software controlling an industrial apparatus, improved data-compression or encryption architecture, latency reduction in a distributed system, or hardware-specific model optimisation. Inventions are at higher risk under Section 3(k) where the contribution is limited to financial rules, business-process workflows, SaaS platform logic, recommendation engines, or a machine-learning model described without a specific technical application.
How Intepat handles software and AI patents
Engagement opens with a technical briefing in which the Patent Agents review the invention disclosure to identify the technical effect and distinguish it from the underlying algorithm or business logic. Where the technical effect is clear, drafting proceeds; where the technical foundation needs strengthening, the Agent works with the inventors to identify the best technical-effect argument the invention supports before drafting begins.
Specification drafting prioritises claim scope within what Indian examination standards will sustain: broad enough to capture the inventive concept, specific enough to anchor a technical effect that will survive a Section 3(k) objection. Independent claims are reviewed against the CRI Guidelines framework before filing.
On receiving an FER citing a Section 3(k) objection, the Patent Agent prepares a substantive response: technical-effect arguments referenced to specific paragraphs of the description, prior-art distinctions where novelty or inventive step are also challenged, and claim amendments where the specification supports a narrower but defensible scope.
For inventions with international filing intent, the Agent coordinates claim strategy across target jurisdictions, including through a PCT international application where appropriate. Indian claim structure diverges from USPTO and EPO practice in how technical character is framed; coordinated drafting from the outset reduces the cost of reconciling divergent claim sets at each national phase.
Documents and inputs required
Invention disclosure: a description of what the software or AI system does, the technical problem it addresses, and how the solution differs from prior approaches
Schematic diagrams or flowcharts illustrating the system architecture or process steps
Benchmark data, test results, accuracy improvements, latency reduction, memory savings, or other technical performance indicators where available
Details of any planned disclosure, product launch, investor presentation, academic publication, or public code release
Prior-art awareness: publications, open-source code, or products the inventors are aware of
Inventor details and assignment authorisation for corporate applicants
For PCT or foreign-priority filings: copies of earlier-filed applications and priority documents
Common pitfalls Intepat protects against
Claiming the algorithm rather than the technical solution.
Claims describing the mathematical model or program steps without anchoring them in a specific technical outcome are the most common source of Section 3(k) rejection.
Inadequate description of the technical effect.
A well-framed claim is undermined by a specification that does not substantiate the technical effect with sufficient detail. Quantified performance data provides a measurable basis for the technical-effect argument at the FER response stage.
Narrowing claims unnecessarily on first response.
A Section 3(k) objection does not automatically require amendment. A substantive technical-effect argument may preserve broader claim scope; in other cases, targeted amendments may be more effective.
Ignoring PCT claim-drafting implications.
Claims optimised solely for Indian filing often create difficulties at the USPTO or EPO national phase. Where international protection is intended, coordinated claim strategy from the outset is more cost-effective than re-drafting at each national phase.
Conflating Section 3(k) with business-method exclusions.
An AI system automating a business process faces two separate exclusion risks: the algorithm per se and the business method per se. Both must be addressed in the specification and claims independently.
Who this is for
Related IP services
Software and AI Patent FAQs
Can software be patented in India?
Yes, when the claimed invention is not merely a computer programme per se and demonstrates a technical effect or technical advancement. Section 3(k) excludes a computer programme per se, but "per se" limits the exclusion: an invention is not rejected solely because it is implemented in code. Improvements in processing efficiency, novel hardware-software architectures, and AI systems producing specific physical outcomes have been protected. Pure algorithmic methods and mathematical optimisations remain excluded.
What is the test for technical effect under the CRI Guidelines?
The CRI Guidelines direct examiners to assess whether the claimed contribution is technical in character. A technical effect is demonstrated when the invention produces an outcome specific to a particular technical problem, concrete and observable in a physical or computational environment, and not already present in the prior art. Examples recognised in Indian practice include improvements in memory utilisation, faster data-transmission protocols, and AI systems producing improved physical-sensing accuracy. The test is applied to the claim as a whole.
How are claims drafted to navigate Section 3(k)?
The principal technique is to anchor independent claims in the technical effect rather than the software steps in isolation. Method claims describe the technical process and outcome; apparatus claims describe a specific configured system rather than a generic computer running a programme. The description supports the technical-effect argument with architecture diagrams, performance data, and explicit identification of what the invention improves over the prior art.
Is an AI model itself patentable in India?
A model described purely as a mathematical structure, a set of weights, or a learning algorithm is vulnerable to rejection under Section 3(k). A specific technical application may be patentable: a model deployed to solve a defined technical problem in a real-world system, such as anomaly detection in industrial sensors or network-latency reduction, is assessed on whether the combination produces a technical effect. The specification must identify the technical problem, the model's role, and the measurable outcome.
Should I file before launching or publishing the software?
Yes. In India, patent applications must be filed before any public disclosure. A product launch, conference presentation, investor demo, academic publication, or public code repository may destroy novelty. Filing a provisional establishes a priority date with minimal disclosure and preserves the right to file a complete specification within 12 months.
What is the typical examination experience for AI and ML inventions?
AI and machine-learning applications frequently receive Section 3(k) objections at the FER stage, particularly where claims focus on a model or learning algorithm without clearly reciting the technical problem and effect. The response identifies the specific technical problem the AI system solves, the measurable effect produced, and how the architecture contributes beyond a generic machine-learning model. Where the technical effect is well-documented in the specification, many such objections can be addressed through substantive argument.
How do Indian, USPTO, and EPO approaches differ?
India requires a demonstrable technical effect under the CRI Guidelines and Section 3(k). The USPTO applies the Alice/Mayo framework, excluding abstract ideas on a computer unless an inventive concept integrates the idea into a practical application. The EPO requires technical character under Article 52 EPC. The three share a focus on technical character but differ in how it must be articulated; coordinated drafting addresses all three.
Are computer-implemented mechanical inventions treated as software patents?
Where software controls a mechanical or industrial process and is integral to a novel mechanical outcome (such as a manufacturing process optimised by a machine-learning algorithm), the claim is typically examined for novelty and inventive step rather than refused at the Section 3(k) threshold. Intepat assesses borderline cases at the pre-drafting stage.
Outline Your Software and AI Patent Requirements
Send your details and a Patent Agent will respond within one business day. All consultations are confidential.