SR&ED Claim Best Practices: Avoid Common Pitfalls
SR&ED claims get rejected or reduced more often than they need to be. The CRA's criteria are specific, and small mistakes can cost you thousands. Here's how to avoid the most common pitfalls and maximize your claim.
Technical Uncertainty: The Core Criterion
SR&ED requires that you faced technical uncertainty—a problem that couldn't be resolved by a competent professional using existing knowledge. For software and AI work, this often means:
- Performance or scaling challenges with no known solution
- Integration problems requiring novel approaches
- Unproven algorithms or architectures
Pitfall: Describing routine development as R&D. Adding a standard feature or following a tutorial doesn't qualify. The uncertainty must be genuine.
Documentation: Your Best Defence
The CRA may review your claim years later. Contemporaneous documentation is critical.
Best practices:
- Weekly or bi-weekly technical notes — What did you try? What worked? What didn't?
- Screenshots and logs — Evidence of experiments and iterations
- Time tracking by project — Eligible vs. ineligible hours
Pitfall: Reconstructing documentation at year-end. Auditors can often tell. Document as you go.
Eligible vs. Ineligible Work
Not all R&D-adjacent work qualifies:
| Eligible | Ineligible |
|---|---|
| Experimentation to resolve technical uncertainty | Market research |
| Prototyping and testing | Routine debugging of known issues |
| Integration of novel systems | General business development |
Pitfall: Over-claiming. If you can't articulate the technical uncertainty, leave it out.
Maximizing Your Claim
- Separate R&D from product work — Time allocation should be defensible
- Include contractor costs — Subcontractors doing eligible work can be claimed
- Claim all provinces — Provincial incentives often stack with federal
- File on time — SR&ED claims have strict deadlines (typically 18 months after year-end)
Conclusion
A well-documented, conservative claim that sails through review is better than an aggressive one that gets challenged. When in doubt, focus on the work where the technical uncertainty was clearest.