Why the USPTO Alice/Mayo Framework and the EPO COMVIK Approach Make International Patent Drafting Difficult

CLIENT EDUCATION GUIDE

For software, artificial intelligence, data analytics, fintech, diagnostic, optimization, and mixed business-technical inventions

*Prepared as a general educational overview. Patentability depends on the particular invention, the claims, the disclosure, and the prior art.

Current through September 2026

Executive Summary

For computer-implemented inventions, the United States and Europe often examine the same technology through different legal lenses. A patent application intended for both jurisdictions must therefore do more than describe a commercially valuable idea. It should identify the technical problem, the technical mechanism used to solve that problem, and the technical effect that follows from the claimed mechanism.

The central drafting challenge

The United States asks whether the claim is impermissibly directed to an abstract idea or other judicial exception. The EPO may accept that a computer-implemented claim has technical character, but under COMVIK it may refuse to give inventive-step weight to features that do not contribute to a technical effect. The safest specification is drafted from the outset to support both analyses.

Six points clients should understand

A commercially clever idea is not automatically a patentable technical invention. A pricing rule, business workflow, recommendation, classification, or prediction can be valuable yet legally vulnerable if the application does not identify the technological substance.
Adding a generic computer usually does not solve the problem. In the United States, generic computer implementation may leave an abstract idea abstract. At the EPO, a computer may clear the initial Article 52 hurdle, while the non-technical features still fail to support inventive step.
The specification must explain how the improvement occurs. Words such as efficient, optimized, intelligent, or improved are conclusions. The application should explain the mechanism that causes the improvement and, where possible, how the improvement can be measured.
Claims and description must work together. A detailed specification is not enough if the claim omits the features that produce the asserted technological improvement. Conversely, a claim amendment years later may be impossible if the needed technical feature was never disclosed.
Europe can be especially unforgiving about later-added technical detail. Article 123(2) EPC generally bars adding subject matter that is not directly and unambiguously derivable from the application as filed. A missing technical linkage often cannot simply be supplied after examination begins.
The application should preserve a ladder of claim scope. The broadest claim should capture the inventive architecture, while dependent claims and alternative embodiments preserve progressively more specific technical fallback positions for prosecution in different jurisdictions.

Practical consequence

For software, AI, data-processing, and mixed business-technical inventions, a substantial part of the attorney’s work occurs before drafting the claims. The attorney must identify which aspects are commercially important, which aspects are genuinely technical, which technical effects can be supported, and which fallback positions should be preserved before the filing date.

1. Two Legal Frameworks Applied to the Same Invention

The frameworks overlap in their practical concern with technological substance, but they are not the same test. Treating them as interchangeable can produce an application that is well positioned in one jurisdiction and unexpectedly weak in the other.

UNITED STATES: ALICE/MAYO

1. Is the claim within a statutory category?

2. Does it recite a judicial exception, commonly an abstract idea?

3. If so, is the exception integrated into a practical application, including through an improvement to computer functionality or another technology?

4. If still directed to the exception, do additional elements amount to significantly more?

5. Separate inquiries under Sections 102, 103, and 112 still apply.

EUROPE: COMVIK

1. Does the claimed subject matter have technical character under Article 52?

2. For a mixed claim, which features actually contribute to technical character?

3. Which differences from the prior art produce a technical effect?

4. Non-technical requirements may be treated as constraints given to the skilled person.

5. Inventive step under Article 56 must rest on the technical contribution.

The simplest useful distinction

As a shorthand, the United States often asks whether the claim is too abstract, while the EPO often asks which parts of the claim actually count as a technical contribution. That shorthand is not a complete statement of either body of law, but it explains much of the drafting difficulty.

Question	USPTO	EPO
Primary concern	Subject matter eligibility under 35 U.S.C. Section 101.	Inventive step for mixed technical and non-technical claims under Article 56 EPC.
Generic computer	Usually does not by itself convert an abstract idea into eligible subject matter.	Often supplies technical character for a computer-implemented method, but ordinarily does not make non-technical features inventive.
Business rules	May fall within the abstract-idea category of certain methods of organizing human activity, particularly for economic or commercial rules.	Generally non-technical and cannot, by themselves, support inventive step.
Mathematics and AI	May trigger abstract-idea analysis, but a claimed technological improvement can support eligibility.	Mathematical models are abstract in isolation, but may contribute where they serve a technical purpose or specific technical implementation.
Best drafting asset	A concrete explanation of the technological improvement and claim language reflecting it.	A clear technical problem, technical mechanism, and technical effect disclosed in the application as filed.

2. What the USPTO Is Asking Under Alice/Mayo

Section 101 identifies broad statutory categories of patentable subject matter, but Supreme Court decisions recognize judicial exceptions for abstract ideas, laws of nature, and natural phenomena. For software and data-processing inventions, the recurring issue is the abstract idea exception.

Step 2A: Is the claim directed to a judicial exception?

The USPTO asks first whether the claim recites a judicial exception and, if so, whether the claim as a whole integrates that exception into a practical application. For many computer-implemented inventions, the key question is whether the claimed arrangement improves the functioning of a computer or another technology, rather than merely using a computer as a tool to carry out an abstract task.

Step 2B: Is there significantly more?

If the claim remains directed to the judicial exception, the analysis asks whether the additional elements, individually and in combination, provide an inventive concept amounting to significantly more than the exception itself. Merely adding conventional computing components to an otherwise abstract business or mathematical process may not be enough.

Why the specification matters in U.S. practice

Current USPTO guidance directs examiners to consider asserted improvements to computer functionality or another technical field. If the applicant relies on a technological improvement, the specification should explain how the invention produces that improvement, and the claim should reflect the features that provide it.

The significance of Ex parte Desjardins

In the 2025 precedential Appeals Review Panel decision Ex parte Desjardins, the USPTO found patent eligibility where the claims reflected an improvement in how a machine-learning model itself operated. The specification described benefits including reduced storage, reduced system complexity, and learning new tasks while protecting knowledge about earlier tasks. The important drafting lesson is not that any machine-learning claim is eligible. It is that the disclosed technological improvement and the claim language must correspond.

3. What the EPO Is Asking Under COMVIK

The EPO approaches computer-implemented inventions differently. A computer-implemented method using technical means will ordinarily have technical character for purposes of Article 52. The difficult question often comes later, when inventive step is assessed under Article 56.

Mixed claims are divided conceptually into contributing and non-contributing features

Under COMVIK, all features that contribute to the technical character of the invention are considered in the inventive-step analysis. Features that do not make such a contribution cannot establish inventive step. A feature can therefore be novel, commercially important, and expressly recited in the claim, yet provide little help if it solves only a non-technical problem.

A new business rule can be treated as a requirement, not as the invention

Suppose a claim contains a new pricing rule implemented by conventional servers. If the pricing rule is regarded as non-technical, the EPO may treat the rule as a requirement supplied to the technically skilled person and ask whether the technical implementation of that requirement was inventive. If the implementation is conventional, novelty in the business rule itself may not save the claim.

The COMVIK danger

A difference from the prior art does not automatically count toward inventive step. The critical question is whether that difference contributes to a technical effect serving a technical purpose.

Artificial intelligence and machine learning

The EPO Guidelines treat AI and machine-learning models as mathematical in nature when considered by themselves. They can nevertheless contribute to technical character when they provide a technical solution to a technical problem, for example through application in a technical field or adaptation to a specific technical implementation. The EPO gives technical examples such as identifying irregular heartbeats in a heart-monitoring apparatus or classifying low-level image, video, audio, or speech features. By contrast, classification based only on linguistic meaning is ordinarily non-technical.

4. Why One International Application Is Difficult to Draft

The attorney is not simply trying to satisfy two sets of wording preferences. The application must preserve legally useful ways of characterizing the same invention under two different frameworks.

Example: AI monitoring of industrial equipment

A commercial description might say, “AI predicts machine failure.” That statement may be accurate, but it is not a sufficient patent disclosure. A stronger application may explain that the system receives vibration signals from specified sensors, preprocesses those signals, generates defined representations, selectively weights portions of the sensor information, applies a particular model architecture, updates the model under defined conditions, produces a predicted mechanical condition, and modifies machine operation in response.

The same disclosure may then support different arguments. In the United States, the claim may be characterized as an improvement in machine monitoring or control rather than a mathematical prediction in the abstract. At the EPO, the data analysis may be relied upon as part of a technical solution because physical sensor signals are processed to determine and control a technical operating state.

The result is not the invention

Inventors often describe objectives: “detect fraud,” “predict disease,” “optimize routing,” or “identify the best result.” Examiners can respond with a single question: How? The application should therefore move from the desired result to the mechanism that produces it.

A useful drafting chain

Technical problem -> specific technical mechanism -> resulting technical effect -> claim language that reflects the mechanism and effect.

5. The Technical Effect and the Causal Link

Both systems reward substance over labels. It is rarely helpful simply to state that a system is efficient, intelligent, secure, or optimized. The application should identify what feature causes the improvement and why.

Weak statement

Conclusion only

“The invention improves network efficiency.”

Stronger technical narrative

Mechanism and effect

Conventional systems repeatedly transmit complete sensor datasets to a central server. The invention generates defined feature vectors at an edge processor and transmits only selected feature vectors satisfying specified conditions. The architecture therefore reduces communication bandwidth while preserving information required for remote fault analysis.

The second explanation gives the examiner a technical story to evaluate. It identifies the prior technical burden, the mechanism that changes system operation, and the technical consequence.

Evidence can help

Numerical evidence is not always required, but it can materially strengthen the disclosure when available. Useful measures may include processor cycles, memory usage, latency, bandwidth, power consumption, signal-to-noise ratio, error rate in a technical detection process, synchronization accuracy, image quality, model-storage requirements, or control stability. For AI and machine-learning inventions, the EPO Guidelines expressly state that a claimed technical effect may be apparent from the disclosure or established by explanation, mathematical proof, experimental data, or similar evidence. Mere assertion may not be enough.

6. The Breadth Paradox

Patent applicants want broad claims because broad claims are harder to design around. Yet a claim can become so broad that the technical mechanism disappears and only the desired result remains. That can make the claim more vulnerable under both Alice/Mayo and COMVIK.

A claim stating only “analyzing data to determine an operating condition” may be commercially broad but legally fragile. A claim limited to one exact sensor, one transform, one threshold, and one processor may be easier to defend but commercially too narrow. Good drafting attempts to preserve a hierarchy between those extremes.

A practical claim ladder

Broad independent claim: captures the inventive technical architecture without unnecessary implementation detail.
First fallback: adds the data source, signal type, system interaction, or processing relationship that anchors the invention technically.
Second fallback: adds a particular data structure, model operation, control function, network arrangement, or resource-saving mechanism.
Further fallbacks: add implementation details, measurement thresholds, particular hardware relationships, training constraints, or other narrower features that remain commercially meaningful.

7. Why the Original Filing Must Anticipate Future Amendments

A major danger in international drafting is assuming that missing technical detail can be supplied later when an examiner raises an eligibility or inventive-step objection. The EPO is particularly strict. Under Article 123(2), an amendment generally cannot add subject matter that the skilled person could not derive directly and unambiguously from the application as filed.

Suppose an application says only that an algorithm “efficiently processes the data.” During European examination, the applicant later explains that the algorithm reduces memory use because it stores only a compressed intermediate representation. That may be an excellent technical argument. If the original application did not disclose the compressed representation or its relationship to memory consumption, however, the applicant may be unable to add the feature to the claim.

Drafting implication

Technical alternatives, causal relationships, and fallback embodiments should be disclosed before filing, even when they are not expected to appear in the first independent claim. The purpose is to preserve future prosecution options without adding new matter later.

U.S. practice creates a related problem through the written-description and new-matter requirements. A later claim amendment still needs support in the application as originally filed. The safest approach is therefore the same: preserve technical detail early.

8. Three Examples Showing Why Context Matters

Example A: Dynamic pricing

A new method for choosing hotel-room prices may be commercially innovative. If the novelty lies only in economic pricing rules, both jurisdictions can be difficult. The United States may view the claim as an abstract commercial practice. The EPO may treat the pricing rules as non-technical. If, however, the invention also uses a new distributed architecture that synchronizes pricing updates among reservation servers while reducing network traffic, the technical architecture may provide a substantially stronger patent position.

Example B: AI machine monitoring

“Use AI to predict bearing failure” is result-oriented. A stronger disclosure explains the sensors, signal conditioning, feature extraction, model operation, training or update rules, fault classification, and resulting machine-control action. The invention can then be framed as industrial monitoring and control, not merely as mathematical prediction.

Example C: Classification

A model that classifies legal documents by subject matter may primarily serve a linguistic or cognitive purpose. A similar mathematical technique applied to packet-level telemetry to identify corrupted communications and automatically alter network routing may have a very different patent analysis because the classification is tied to operation of a technical system.

9. How We Draft for Both Jurisdictions

A robust application is usually developed around several complementary narratives, all grounded in the actual invention. The objective is not to insert the word “technical” repeatedly. The objective is to capture the technological substance at enough levels of detail to remain useful during later prosecution.

Drafting inquiry	Purpose
Commercial objective	What useful result does the product or service provide?
Technical problem	What technological limitation, bottleneck, instability, resource burden, measurement problem, control problem, or computing problem had to be solved?
Technical mechanism	What specific system architecture, data flow, algorithmic relationship, model operation, signal-processing sequence, control loop, data structure, or hardware interaction solves that problem?
Technical effect	What changes because of that mechanism: less memory, less bandwidth, faster processing, better signal quality, more stable control, improved machine operation, reduced power, improved security, or another technical result?
Alternatives	What other implementations achieve the same technical purpose without tying the claims to one commercial embodiment?
Claim ladder	Which features belong in the broadest claim, and which should be preserved as progressively narrower fallback positions?

Why inventor interviews matter

For these inventions, the engineer who implemented the system may know facts that the business-side inventor does not initially recognize as patent-significant. Edge processing, selective data transmission, memory structures, recalibration logic, model-update rules, synchronization, security measures, and other implementation details can become central to patentability. The drafting process therefore benefits from both the commercial story and the engineering story.

10. What Clients Should Expect

A high-quality application involving software, AI, or data processing may require more preparation than its apparent simplicity suggests. Much of the attorney’s work is directed to identifying the legally relevant invention before the final claims are written.

Expect questions that go beyond what the product does and focus on how it works.
Expect requests for architecture diagrams, data-flow diagrams, examples, benchmarks, and alternative implementations.
Expect the drafting attorney to distinguish business rules from technical mechanisms rather than treating them as interchangeable.
Expect the application to disclose more detail than appears in the broadest claim so that later fallback positions remain available.
Do not assume that a missing technical explanation can be added after filing. In Europe, that assumption can be especially damaging.

Bottom line for clients

Alice/Mayo and COMVIK are not merely prosecution doctrines that an attorney can address years after filing. They are drafting constraints. The strongest time to identify the technical problem, mechanism, and effect is before the patent application is filed.

Inventor Questionnaire: Information We Need Before Drafting a Software, AI, or Data-Driven Patent Application

This questionnaire is intended to identify the technical facts that may be important under both U.S. and European practice. Short answers are acceptable. Attach diagrams, source-code excerpts, architecture documents, test results, product requirements, or other materials where they answer the question more efficiently.

Please involve the right people

Where possible, have at least one person who understands the commercial objective and one person who understands the actual technical implementation participate in the invention disclosure process.

1. In one or two sentences, what does the invention do?

Describe the product or process in ordinary language. What problem does the customer or user perceive?

____________________________________________________________________________________________________________

2. What is new compared with the conventional approach?

Identify the feature, relationship, architecture, process, or algorithm that you believe competitors do not use.

____________________________________________________________________________________________________________

3. What technical problem had to be solved?

Examples include memory consumption, bandwidth, latency, signal noise, synchronization, model drift, processor load, power consumption, control instability, sensor error, network congestion, data integrity, security, or hardware constraints.

____________________________________________________________________________________________________________

4. How did prior systems handle that technical problem?

Describe the previous architecture or process and its technical weakness. If there is no known prior solution, explain what was previously impossible or impractical.

____________________________________________________________________________________________________________

5. What specific mechanism solves the technical problem?

Describe the actual steps, data structures, processing relationships, system components, control logic, model operations, or hardware interactions. Avoid describing only the desired result.

____________________________________________________________________________________________________________

6. What technical effect or measurable improvement results?

For example: lower memory usage, reduced network traffic, lower latency, improved accuracy of a technical measurement, reduced processor load, improved signal quality, reduced power consumption, improved machine control, or improved security.

____________________________________________________________________________________________________________

7. Do you have measurements, benchmarks, simulations, or test results supporting that effect?

If yes, identify them and attach available results. Approximate internal measurements can still be useful during drafting.

____________________________________________________________________________________________________________

8. What are the principal inputs to the system?

Identify sensors, images, audio, signals, transaction streams, databases, user inputs, machine states, or other data sources. Explain how the data are obtained.

____________________________________________________________________________________________________________

9. What processing is performed on those inputs?

Describe preprocessing, filtering, normalization, feature generation, transformations, mathematical relationships, rules, model inference, model training, ranking, selection, optimization, or other operations.

____________________________________________________________________________________________________________

10. What outputs are produced, and what happens because of them?

Does the system merely present information, or does it alter routing, machine operation, treatment apparatus operation, storage, network behavior, model behavior, security state, or another technical process?

____________________________________________________________________________________________________________

11. What computing architecture is used?

Identify edge devices, mobile devices, cloud servers, local servers, distributed nodes, processors, accelerators, memory, databases, networks, sensors, and interfaces. Explain which operations occur where.

____________________________________________________________________________________________________________

12. Are there technical reasons the operations are divided among different devices or locations?

Examples include latency, privacy, bandwidth, processor capacity, resilience, power, synchronization, or security.

____________________________________________________________________________________________________________

13. If AI or machine learning is involved, what exactly is the model doing?

Identify the model type if relevant, the input representation, output, training objective, update method, inference process, and any model-specific improvement. Explain whether the novelty lies in the model itself, its training, its deployment, its data, or its use in a technical process.

____________________________________________________________________________________________________________

14. If training data are important, what characteristics of the training data matter?

Describe data sources, labels, sampling, preprocessing, feature selection, synthetic data, temporal relationships, class balance, sensor conditions, or other characteristics needed to obtain the claimed result.

____________________________________________________________________________________________________________

15. Does the invention improve the computer or model itself?

Examples include reduced storage, improved data structures, reduced computational complexity, better continual learning, reduced catastrophic forgetting, faster inference, improved memory allocation, improved load balancing, or lower communication overhead.

____________________________________________________________________________________________________________

16. Does the invention control or improve a physical or technical system?

Identify any machine, sensor, network, medical apparatus, manufacturing process, communications system, imaging system, control system, or other technology affected by the output.

____________________________________________________________________________________________________________

17. Which aspects are business, administrative, financial, or organizational rules?

Separating these from the technical implementation helps identify which features may receive less weight under EPO COMVIK analysis.

____________________________________________________________________________________________________________

18. Could the same commercial objective be implemented through a different technical architecture?

Describe alternative sensors, processors, model types, network arrangements, data structures, control techniques, or processing sequences that should remain within the invention.

____________________________________________________________________________________________________________

19. Which implementation details are required, and which are merely preferred?

This distinction helps us avoid unnecessarily narrow claims while preserving useful fallback positions.

____________________________________________________________________________________________________________

20. What are the likely design-around strategies?

How might a competitor obtain the same practical benefit while changing the architecture, data source, model, processing location, or sequence of steps?

____________________________________________________________________________________________________________

21. What prior patents, products, publications, open-source projects, or internal systems are closest?

Provide names, links, patent numbers, papers, screenshots, or other materials. Also identify what each reference lacks.

____________________________________________________________________________________________________________

22. Has any version of the invention been publicly disclosed, offered for sale, used commercially, demonstrated, published, or shared outside confidentiality obligations?

Provide dates and circumstances. Timing can affect patent rights in different countries.

____________________________________________________________________________________________________________

23. Who contributed to the technical conception of the claimed subject matter?

List the people who conceived the technical features, not merely those who implemented instructions. Inventorship is a claim-specific legal determination.

____________________________________________________________________________________________________________

24. What documents should we review before drafting?

Examples: architecture diagrams, flowcharts, source-code excerpts, model diagrams, notebooks, technical specifications, test results, product requirements, white papers, presentations, or internal invention disclosures.

____________________________________________________________________________________________________________

Helpful materials checklist

[ ] System architecture diagram
[ ] Data-flow diagram
[ ] Flowchart of the principal algorithm or process
[ ] AI or machine-learning model diagram
[ ] Example inputs and outputs
[ ] Benchmarks or test results
[ ] Description of conventional approach and its shortcomings
[ ] Alternative embodiments and design-around ideas
[ ] Relevant patents, papers, products, or open-source references
[ ] Dates of any public disclosures, sales, demonstrations, or releases

Selected Authorities and Further Reading

The following primary sources are useful starting points for the legal frameworks summarized in this guide. The discussion above is an independently written summary and paraphrase of these authorities and does not reproduce extended passages from them. The sources are included for reference rather than as a substitute for advice concerning a particular invention.

USPTO MPEP Section 2106, Patent Subject Matter Eligibility: https://www.uspto.gov/web/offices/pac/mpep/s2106.html

USPTO, Ex parte Desjardins, Appeal No. 2024-000567, precedential designation and summary: https://www.uspto.gov/subscription-center/2025/ptab-designates-precedential-appeals-review-panel-decision

USPTO, December 5, 2025 subject matter eligibility guidance update following Ex parte Desjardins: https://www.uspto.gov/subscription-center/2025/uspto-updates-subject-matter-eligibility-guidance-mpep

EPO Guidelines G-VII, 5.4, Claims comprising technical and non-technical features, COMVIK: https://www.epo.org/en/legal/guidelines-epc/2026/g_vii_5_4.html

EPO Guidelines G-II, 3.3.1, Artificial intelligence and machine learning: https://www.epo.org/en/legal/guidelines-epc/2026/g_ii_3_3_1.html

EPO Guidelines G-II, 3.6 and 3.6.1, Programs for computers and further technical effects: https://www.epo.org/en/legal/guidelines-epc/2026/g_ii_3_6.html

EPO Guidelines H-IV, 2.2, Article 123(2) and content of the application as filed: https://www.epo.org/en/legal/guidelines-epc/2026/h_iv_2_2.html

Core cases and doctrine

Alice Corp. Pty. Ltd. v. CLS Bank International, 573 U.S. 208 (2014).
Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012).
Enfish, LLC v. Microsoft Corp., 822 F.3d 1327 (Fed. Cir. 2016).
EPO Board of Appeal T 641/00, Two identities/COMVIK.
EPO Enlarged Board of Appeal G 1/19, Pedestrian simulation.

Important notice

This guide provides general educational information and does not address the patentability, validity, enforceability, or filing strategy for any particular invention. It focuses on computer-implemented subject matter and does not attempt to cover every additional doctrine that may apply to medical, diagnostic, or life-science inventions. U.S. subject matter eligibility law and European computer-implemented invention practice continue to develop. Specific advice should be based on the actual disclosure, claims, prior art, filing history, and intended jurisdictions.

PROTECTING SOFTWARE, AI, AND DATA-DRIVEN INVENTIONS

Executive Summary

Six points clients should understand

1. Two Legal Frameworks Applied to the Same Invention

The simplest useful distinction

2. What the USPTO Is Asking Under Alice/Mayo

Step 2A: Is the claim directed to a judicial exception?

Step 2B: Is there significantly more?

The significance of Ex parte Desjardins

3. What the EPO Is Asking Under COMVIK

Mixed claims are divided conceptually into contributing and non-contributing features

A new business rule can be treated as a requirement, not as the invention

Artificial intelligence and machine learning

4. Why One International Application Is Difficult to Draft

Example: AI monitoring of industrial equipment

The result is not the invention

5. The Technical Effect and the Causal Link

Weak statement

Stronger technical narrative

Evidence can help

6. The Breadth Paradox

A practical claim ladder

7. Why the Original Filing Must Anticipate Future Amendments

8. Three Examples Showing Why Context Matters

Example A: Dynamic pricing

Example B: AI machine monitoring

Example C: Classification

9. How We Draft for Both Jurisdictions

Why inventor interviews matter

10. What Clients Should Expect

Inventor Questionnaire: Information We Need Before Drafting a Software, AI, or Data-Driven Patent Application

Helpful materials checklist

Selected Authorities and Further Reading

Core cases and doctrine

Categories

PROTECTING SOFTWARE, AI, AND DATA-DRIVEN INVENTIONS

Executive Summary

Six points clients should understand

1. Two Legal Frameworks Applied to the Same Invention

The simplest useful distinction

2. What the USPTO Is Asking Under Alice/Mayo

Step 2A: Is the claim directed to a judicial exception?

Step 2B: Is there significantly more?

The significance of Ex parte Desjardins

3. What the EPO Is Asking Under COMVIK

Mixed claims are divided conceptually into contributing and non-contributing features

A new business rule can be treated as a requirement, not as the invention

Artificial intelligence and machine learning

4. Why One International Application Is Difficult to Draft

Example: AI monitoring of industrial equipment

The result is not the invention

5. The Technical Effect and the Causal Link

Weak statement

Stronger technical narrative

Evidence can help

6. The Breadth Paradox

A practical claim ladder

7. Why the Original Filing Must Anticipate Future Amendments

8. Three Examples Showing Why Context Matters

Example A: Dynamic pricing

Example B: AI machine monitoring

Example C: Classification

9. How We Draft for Both Jurisdictions

Why inventor interviews matter

10. What Clients Should Expect

Inventor Questionnaire: Information We Need Before Drafting a Software, AI, or Data-Driven Patent Application

Helpful materials checklist

Selected Authorities and Further Reading

Core cases and doctrine

Related:

Categories