The Programme Most Canadian AI Companies Underuse
NRC IRAP (the Industrial Research Assistance Programme) is the single most valuable funding source for Canadian technology companies, and the majority of Ontario companies developing AI never apply. Some don't know about it. Some assume they won't qualify. Some tried once, got rejected, and gave up.
That's leaving money on the table. IRAP provides non-repayable contributions (grants, not loans) covering up to 80% of eligible project costs. For a typical AI development project with $250,000 in labour costs, that's $200,000 in funding. We've helped Ontario companies secure IRAP funding for AI projects ranging from $50,000 to $1.2 million.
Here's what you need to know.
Exploring IRAP for your AI project? Talk to one of our engineers — we've guided multiple Canadian companies through the IRAP application process for AI initiatives and can tell you quickly whether your project is likely to qualify.
Who Is Eligible
IRAP eligibility comes down to four criteria. All four must be met.
1. Canadian corporation. You must be incorporated in Canada. Foreign-owned subsidiaries with a Canadian incorporation qualify. Sole proprietorships and partnerships do not.
2. Small or medium enterprise. Fewer than 500 full-time equivalent employees. Most Ontario AI companies fall well within this threshold.
3. Profit-oriented. Not-for-profits, universities, and government agencies don't qualify (they have other NRC programmes). You must be a for-profit business.
4. Technology-driven project. The project must involve technological innovation — developing or improving a product, process, or service through R&D. This is where AI projects fit naturally.
One common misconception: you don't need to be a "technology company." A food manufacturer developing AI-based quality inspection qualifies. A logistics company building a demand forecasting model qualifies. A construction firm implementing IoT-based structural monitoring qualifies. The requirement is that the project involves technology development, not that the company is in the technology sector.
What AI Activities Qualify
IRAP funds R&D, not deployment of existing technology. The distinction matters.
Typically qualifies:
- Developing a new ML model for a specific industrial application
- Creating proprietary training datasets or data pipelines
- Building novel feature engineering approaches for domain-specific problems
- Integrating AI with IoT, edge computing, or industrial control systems
- Developing AI-powered products or SaaS platforms
- Adapting existing AI techniques to a new industry or application in Canada
Typically does NOT qualify:
- Buying an off-the-shelf AI platform and configuring it
- Using pre-trained models (ChatGPT, etc.) without modification
- Routine data analysis or business intelligence work
- Deploying a previously developed model to production (though the development phase qualifies)
The key question NRC IRAP asks: "Is there technological uncertainty?" If you don't know whether your approach will work — if there's genuine R&D risk — that's what IRAP funds. If you're following a well-understood recipe, it's not R&D, and IRAP won't fund it.
How Much Funding
IRAP contributions vary based on project scope, but here are realistic ranges for AI projects in Ontario:
- Youth employment: Up to $30,000 per position (50% of salary for new graduates hired onto the project)
- Advisory services: Free access to NRC IRAP Industrial Technology Advisors (ITAs) — experienced technical and business advisors assigned to your company
- Project funding: Typically $50,000 to $1,000,000 in non-repayable contributions
- Cost sharing: IRAP typically covers 50-80% of eligible costs, depending on project stage and company maturity
Eligible costs include:
- Salaries and wages of employees working on the project (the largest component)
- Contractor/subcontractor costs (including AI consulting firms — this is relevant if you're working with us)
- Materials and supplies consumed during R&D
- Equipment rental for the project period
Not eligible:
- Capital equipment purchases (you can lease, not buy)
- Overhead and administrative costs
- Marketing and sales activities
- Costs incurred before the project is approved
That last point is critical. IRAP does not fund retroactively. Costs incurred before your project is approved by NRC are not eligible. Start your application before you start the work.
The Application Process
Step 1: Initial Contact With an ITA
IRAP doesn't have an open application form. The process starts with a conversation with an Industrial Technology Advisor (ITA) in your region. For Ontario companies, contact the NRC IRAP Ontario regional office. Your ITA will assess whether your company and project are a good fit.
Timeline: 1-2 weeks to get an initial meeting.
Step 2: Company Assessment
Your ITA evaluates your company's capacity to execute the project: management team, technical team, financial health, market opportunity. IRAP invests in companies, not just projects. They want to see that a successful project will lead to commercialisation and growth in Canada.
Timeline: 2-4 weeks.
Step 3: Project Proposal
Working with your ITA, you develop a formal project proposal covering:
- Technical objectives and approach
- Work plan with milestones and deliverables
- Budget breakdown by cost category
- Market opportunity and commercialisation plan
- Team qualifications
- Risk assessment
For AI projects, the technical section should clearly articulate the technological uncertainty — what you don't know, what you're trying to figure out, and why existing approaches don't solve the problem.
Timeline: 2-6 weeks depending on proposal complexity.
Step 4: Review and Approval
Your proposal is reviewed by NRC. For projects under $300,000, this is typically handled at the regional level. Larger projects may go to a national review committee.
Timeline: 4-8 weeks for review and decision.
Total Timeline: 10-20 Weeks From First Contact to Approval
Plan accordingly. If you need IRAP funding for an AI project starting in Q3 2026, begin the process now.
SR&ED and IRAP: They're Not Either/Or
A question we get from Ontario companies at least once a month: "Should we apply for IRAP or SR&ED?"
Both. They're complementary, not competing.
IRAP provides upfront funding (cash in hand during the project). SR&ED provides retroactive tax credits (claimed after your fiscal year ends). The same project costs can be supported by both programmes, though you can't claim the IRAP-funded portion of costs under SR&ED (that would be double-dipping).
Here's how it works in practice:
- Your AI project has $300,000 in eligible labour costs
- IRAP covers $200,000 (67%)
- Your company pays $100,000 out of pocket
- At year end, you claim the $100,000 you paid out of pocket under SR&ED
- SR&ED returns approximately $35,000-$45,000 in tax credits (35% for CCPCs on the first $3M of eligible expenditures)
- Your net cost for $300,000 worth of AI development: $55,000-$65,000
That's 78-82% cost recovery. For a Canadian SME developing AI, this is an extraordinary deal. No other G7 country offers this combination of support for technology R&D.
Tips for Approval
Based on our experience helping Ontario companies secure IRAP funding for AI projects:
1. Lead with the business problem, not the technology. ITAs evaluate commercial potential, not just technical merit. "We're building a transformer-based anomaly detection system" is less compelling than "We're reducing unplanned downtime by 40% for Canadian manufacturers, saving each customer $300K/year."
2. Articulate genuine technological uncertainty. "We don't know if our approach will achieve the required 95% accuracy on production data" is good. "We need to implement a well-known algorithm" is not R&D and won't qualify.
3. Have your data house in order. For AI projects specifically, IRAP ITAs will ask about your training data. Where does it come from? How much do you have? Is it labelled? Do you have rights to use it? Answers to these questions signal project maturity.
4. Show a path to revenue. IRAP invests in companies that will grow the Canadian economy. Show how the project leads to a product, service, or capability that generates revenue — either from selling the AI solution or from operational savings that make your company more competitive in the Canadian market.
5. Build a relationship with your ITA. The ITA is your advocate within NRC. They're not gatekeepers — they're advisors who want to help Ontario companies succeed. Be responsive, be transparent about risks, and take their feedback seriously. Many successful IRAP applicants in Canada describe the ITA relationship as more valuable than the funding itself.
6. Don't wait until you need the money. IRAP takes 10-20 weeks from first contact to approval. Companies that approach IRAP under time pressure get worse outcomes than companies that plan ahead.
What Happens After Approval
Once approved, IRAP doesn't hand you a cheque and walk away. The programme includes:
- Regular milestone reviews: Your ITA checks progress against the work plan (typically quarterly)
- Claims processing: You submit cost claims with supporting documentation; IRAP reimburses within 4-6 weeks
- Advisory support: Your ITA remains available for technical and business advice throughout the project
- Network access: IRAP connects you with other NRC programmes, potential customers, and international partners through the Canadian Trade Commissioner Service
Stop Leaving Money on the Table
IRAP exists to help Canadian companies like yours develop technology. The programme has an annual budget exceeding $400 million. If you're developing AI in Ontario and you haven't applied, you're subsidising your competitors who have.
The process isn't difficult. The funding is substantial. And the advisory relationship with your ITA adds value beyond the dollars.
Talk to an engineer about IRAP funding for your AI project. We'll help you assess eligibility, structure the technical narrative, and connect you with the right ITA in Ontario. No charge for the initial consultation.
For a complete framework on moving your AI project from concept to production, read our guide on AI proof of concept to production in 90 days.
Learn more about our AI consulting services and how we help Canadian enterprises build and deploy AI systems.
