← All insights

SR&ED

What CRA Actually Means by Technological Uncertainty

Rejected claims describe a product. Approved claims describe a problem nobody knew how to solve.

7 min readOrientum advisory team

The single biggest predictor of a successful SR&ED claim is not how hard the engineering was. It is whether the write-up describes the work in the language the program is built around: uncertainty, hypothesis, systematic investigation, advancement.

The Distinction That Decides Claims

Building something difficult is not, by itself, eligible. The question is whether a competent professional in the field, using existing knowledge and standard practice, could have known in advance whether the approach would work and how to achieve it. If yes, that is engineering. If no, that is technological uncertainty, and the work done to resolve it is what you claim.

Two descriptions of the same project

Not eligible as written: "We built a real-time recommendation engine for our marketplace." Eligible as written: "Existing collaborative filtering approaches degraded below acceptable latency at our sparsity level. It was not known whether the ranking quality could be preserved under that constraint. We formulated and tested three approaches and measured the tradeoff."

The Five Questions to Answer in Every Write-Up

  1. What technological problem did you face, stated as a limitation rather than a feature request?
  2. Why could it not be resolved with standard practice and publicly available knowledge?
  3. What hypotheses did you form?
  4. What did you do to test them, in what order, and what happened, including the failures?
  5. What technological knowledge did you gain, whether or not the project succeeded?

What Disqualifies Otherwise Good Work

  • Framing the advancement as commercial, faster onboarding, better conversion, rather than technological.
  • Claiming the whole project instead of the portion where uncertainty existed.
  • Routine integration, configuration and debugging described as experimentation.
  • No record of the failed attempts. Failures are evidence of uncertainty, they strengthen a claim rather than weakening it.

What to Do

  1. For each candidate project, write the uncertainty in one sentence before anything else. If you cannot, the project may not qualify.
  2. Separate the eligible portion from surrounding development work, and track hours accordingly.
  3. Capture failed approaches as you go. See the note on documentation habits.

General information for Canadian founders, current to 2026. It is not tax advice and does not account for your specific facts. Rates, thresholds and rules change, confirm the current figures before acting.

Want This Applied to Your Numbers?

A short call is usually enough to tell you whether there is anything here worth acting on.