Written By Louis-Pierre Gravelle

Executive summary

CIPO’s March 2026 practice notice formally abandons the language that drew sustained criticism under PN2020‑04 — notably references to a claim’s “contribution,” a “technological problem,” and a “technological solution.” In their place, the notice re‑centres subject‑matter analysis on purposive construction, explicitly aligning examination practice with recent Federal Court and Federal Court of Appeal jurisprudence.

Yet for software and AI inventions, a closer reading suggests that the practical battleground may not be eligibility doctrine as such, but disclosure quality. By expressly permitting examiners to treat elements described as well‑known, conventional, or disclosed with little detail as common general knowledge (CGK) during purposive construction, the notice introduces a powerful upstream lever — one that may lead to familiar outcomes, albeit through a different pathway.

This article examines the March 2026 practice notice through that lens and considers its implications for software patent drafting and prosecution in Canada.

From “contribution” to purposive construction: a genuine reset

The March 2026 practice notice states unambiguously that references in the MOPOP to a claim’s “contribution,” to a “technological solution to a technological problem,” and to evaluating essentiality based on a “problem and solution” do not apply.

Subject‑matter is instead determined based on purposive construction conducted in accordance with Supreme Court of Canada jurisprudence, before any assessment of patentability.

This restructuring responds directly to concerns expressed by the courts in cases such as Choueifaty, Benjamin Moore, and Dusome, where CIPO’s prior tendency to re‑characterise claims at a later stage — often via an “actual invention” or problem–solution analysis — was found to be inconsistent with the principles of claim construction as set out by the Supreme Court more than 20 years ago.

On its face, this is more than cosmetic. It removes from the formal framework the very concepts that had become lightning rods for judicial review.

The CGK lever: where scrutiny now concentrates

The same practice notice, however, emphasises that in purposive construction:

Elements of the claim that are presented in the specification as being well‑known, used in a conventional way, or upon which little or no detail is disclosed may constitute common general knowledge.

This sentence may prove to be the most important — and under‑examined — part of the notice for software and AI applications.

In practice, treating software implementation elements as CGK can have a decisive effect. Once such elements are discounted as conventional, what remains of the claim may be characterised as an abstract algorithm or a set of rules, prompting the familiar conclusion that the claimed subject‑matter lacks the requisite physicality.

Viewed through this lens, the March 2026 notice does not so much eliminate the risk of “same old” outcomes as relocate decisive scrutiny of the claim earlier in the analysis. The gatekeeping function shifts from an overt “contribution” inquiry to an implicit assessment of whether the implementation is sufficiently described to be treated as part of the invention, rather than as background knowledge.

Physicality remains the ultimate constraint

Unsurprisingly, nothing in the March 2026 practice notice suggests a relaxation of CIPO’s long‑standing position that patentable subject‑matter must have physical existence or manifest a discernible physical effect or change.

For computer‑implemented inventions, the notice confirms that merely reciting a computer is not enough. Physicality may be found where there are additional physical essential elements (such as measurement or sensing), or where the invention improves the functioning of the computer itself.

Where the computer is the only physical element, the notice proposes asking the so‑called “Schlumberger question”: does the claim amount, in substance, to programming a mathematical formula or abstract rules on a computer? If so, some “something more” is required.

The point is not that this approach is new, but that — combined with aggressive CGK filtering — it provides a well‑trodden path to outcomes that applicants will find familiar.

Why “fresh coat of paint” critiques persist

Commentary to date has rightly welcomed the elimination of “actual invention” as a free‑standing inquiry and the formal abandonment of problem–solution language. Those changes matter, particularly for appeal and judicial review posture.

At the same time, it is apparent that practitioners remain skeptical about whether examination outcomes will materially change for software and AI inventions. The March 2026 notice leaves intact both the physicality requirement and the examiner’s ability to narrow the invention during purposive construction by treating lightly described elements as CGK.

In other words, the architecture of the analysis has changed — but the pressure points may not have.

Practical implications: disclosure as an eligibility strategy

A key practical consequence of the March 2026 notice is that disclosure depth now operates as a form of eligibility risk management for software inventions.

Applicants who describe critical implementation steps at a high level, or who characterise them as routine or conventional, may find those elements discounted at the construction stage — with predictable downstream effects on subject‑matter eligibility.

Conversely, detailed disclosure of how claimed software functionality is implemented, interacts with hardware, or improves computer operation may help resist CGK characterisation and preserve a more concrete “nature of the invention.”

This shift raises familiar trade‑offs: more detailed disclosure can strengthen eligibility positioning, but it also increases drafting burden and may expose implementation detail that businesses would prefer to protect as know‑how.

Looking ahead – and inviting debate

The March 2026 practice notice is not merely a stylistic rewrite of PN2020‑04. It represents a conscientious effort by CIPO to realign examination practice with binding jurisprudence and to avoid analytical shortcuts that courts have rejected.

At the same time, for software and AI inventions, the notice places renewed emphasis on an issue that has always mattered — but now does so explicitly: what the specification says, and how it says it.

Whether the new framework leads to genuinely different outcomes in examination, or whether familiar results re‑emerge through CGK‑based construction, will ultimately be an empirical question.

That question is now squarely in the hands of practitioners and examiners alike. Will we be seeing meaningful changes in how software claims are construed and assessed — or simply a re‑labeling of the same objections through a different entry point?

I would welcome debate on that point, particularly from those already prosecuting applications under the March 2026 guidance.