“We developed a revolutionary AI platform. We’re definitely doing SR&ED innovation, right?”
This is probably the sentence we hear most often.
And yet…
The answer isn’t necessarily yes.
Yes, a company can invest several million dollars, mobilize a team of 30 developers for two years, and create a completely unprecedented product… without carrying out a single eligible SR&ED project.
Conversely, an 8-person SME can spend three months solving a very specific problem and carry out genuine experimental research.
Paradoxical?
Not really.
Imagine two teams.
The first develops a new management software. It assemble proven technologies, follows best practices, consults documentation, fixes bugs, and delivers an excellent product.
The work is remarkable.
But, from a SR&ED perspective, there is nothing experimental about it.
The second team seeks to run a real-time artificial intelligence engine despite contradictory constraints on performance, accuracy, and resource consumption.
Scientific articles don’t provide the answer.
Experts don’t know which approach will work.
The initial architectures fail.
The subsequent ones create new problems.
The team formulates hypotheses, experiments, measures, questions its own choices… until discovering an approach that finally works.
Therein lies the difference.
SR&ED does not begin when you start writing code.
It begins when no one knows how to solve the problem yet.
In fact, this is the most common mistake we see.
Many companies think SR&ED is about creating an innovative product.
In reality, it is about generating new technological knowledge.
The final product is often just the visible consequence of a long series of trials, failures, and learnings.
And that is precisely where the value of SR&ED lies.
So, how do you know if your team might have found itself in such a situation?
Ask yourself a few questions:
- 👉 Did you have to test several architectures before finding one capable of meeting your constraints? *
- 👉 Did performance gains degrade security, accuracy, resource consumption, or availability, forcing you to experiment with several trade-offs?
- 👉 Were the results obtained impossible to predict before testing?
- 👉 Did best practices, official documentation, or even experts fail to clearly identify the solution?
- 👉 Did you have to develop several prototypes or run numerous experiments before understanding why certain approaches failed?
If several of these situations sound familiar, it is possible that your team accomplished much more than just software development.
Because most developers don’t realize that they are sometimes doing SR&ED.
Not because they are developing innovative software, but because at certain moments in the project, they find themselves facing a question to which no one, neither their team, nor the technical documentation, nor technology vendors, knows the answer yet.
And that is precisely where genuine scientific research, experimental development, and SR&ED-eligible innovation often begin.
Sources
Government of Canada – Scientific Research and Experimental Development (SR&ED) tax incentives