Most people ask a consultant what they charge before they ask who is going to write it. A claim is two jobs in one document: someone has to describe the technical work accurately, and someone has to put the costs together correctly. Different skills, usually different people.

The technical half is the one to ask about. Whoever writes it has to sit with your engineer or your formulator and follow the conversation properly. People who trained as engineers or scientists tend to be good at this, because they have stood in front of an experiment that did not behave and have had to write up work that did not go as expected.

Read the briefing →  The four questions worth asking before you hire anyone, including the one that tells you more than the rest.

You can have both, and companies do it every year. There is one rule underneath it that surprises people: if government money paid for part of a cost, the tax credit is worked out on the part it did not pay. On a $150,000 contribution the federal credit comes down by roughly $52,500, so about $97,500 of it still stays with you.

Three things follow. Timing, because what matters is when you know the money is coming rather than when it reaches the bank. Disclosure, because every application asks for your other funding and the agreement sets a ceiling. And records, because the same hours get divided between programmes rather than counted twice.

Read the briefing →  Including the planning move that is only available while the budget is still being written.

Canada's five Global Innovation Clusters hold close to $2 billion in non-repayable federal contributions, and most companies never look at them. There is no federal form and no single intake. Each cluster is an independent not-for-profit with its own membership and its own calls for projects, and the money only moves through consortia that include at least one Canadian SME, matched dollar for dollar by industry.

NGen covers advanced manufacturing, Protein Industries covers agri-food, DIGITAL covers digital technology, Scale AI covers AI in supply chains, and the Ocean Supercluster covers the ocean economy. The sector boundaries are wider than the names suggest, and companies routinely rule themselves out on the label rather than the mandate.

A cluster contribution is government assistance, so it reduces qualified SR&ED expenditures dollar for dollar, the same as IRAP. On a collaborative project that becomes a negotiation rather than a calculation, because several partners are doing the work and somebody has to decide whose expenditures are whose before anybody files.

Read the full guide →  What each cluster funds, the route in from membership to project agreement, why proposals fail, and what a contribution costs a claim once the grind is counted.

IRAP is the largest source of direct R&D funding for Canadian small and medium businesses, and the least predictable to apply for. There is no application form and no published deadline. You start with a phone call, and the real gate is whether an Industrial Technology Advisor invites a proposal at all.

The part that is usually left out of the conversation is what it does to your tax credit. An IRAP contribution is government assistance, and government assistance reduces qualified SR&ED expenditures dollar for dollar. On a $150,000 contribution against a $775,000 pool, the federal credit falls by $52,500. The money is still worth taking. It is worth roughly sixty-five cents on the dollar, not a hundred, and anyone presenting it as free money on top of a full claim has not done the arithmetic.

The reduction also applies when you can reasonably expect to receive the contribution, not when the cash arrives, which catches companies that treat it as a timing question.

Read the full guide →  Who qualifies, how the process really runs, why proposals stall, and how to sequence the two programmes so the contribution costs you as little credit as possible.

For taxation years beginning after 25 March 2025, Quebec replaced its provincial SR&ED credit, and seven other incentives, with a single refundable credit: the CRIC, the tax credit for research, innovation, and commercialization. If your company does R&D in Quebec, this is the biggest change to your provincial funding in a generation.

The structure is simple. The CRIC pays 30% on the first $1 million of qualified expenditure, and 20% above that threshold. It is fully refundable, which means it pays out in cash even when no tax is owing. The former regime paid 30% on a larger bracket for some claimants, so the effect varies company by company: smaller claimants generally come out ahead, while some larger R&D groups will see the marginal rate fall.

The most important expansion is what now counts. For the first time, the provincial credit covers the capital cost of property used in R&D, equipment included, alongside salaries and eligible subcontracting. Land and buildings remain excluded. For manufacturers and laboratories that invest in instruments and processing equipment for development work, this opens a category of support Quebec never offered through the old R&D credit.

The federal SR&ED credit is untouched. Your claim now has two distinct layers: the federal T661 claim, prepared as before, and the Quebec CRIC computed under its own rules. The two must be coordinated, because expenditure choices on one side affect the other.

If your fiscal year began after 25 March 2025, your current claim is already under the new rules. It is worth reviewing your project list and planned equipment purchases now, rather than at year-end, so the capital component is captured properly.

British Columbia has introduced a refundable Manufacturing and Processing Investment Tax Credit worth 15% of eligible investment. It is one of the most generous provincial capital credits in the country, and because it is new, many B.C. manufacturers do not yet know it exists.

The credit applies to Class 43 machinery and equipment used to manufacture or process goods in the province, and to qualifying manufacturing or processing buildings, for property acquired from 1 April 2026. It is capped at $2 million per eligible property, and associated corporate groups share the limit. The program is scheduled to run for acquisitions through 2031, with rates stepping down afterward.

Eligibility follows a familiar pattern: the claimant must be a Canadian-controlled private corporation with a permanent establishment in British Columbia, and must not be exempt from B.C. income tax. The credit is claimed through the corporate tax return rather than by separate application.

Two planning points matter. First, timing: property must be acquired and become available for use within the program window, so equipment decisions in the next fiscal year should be checked against the April 2026 start line. Second, stacking: the credit sits alongside federal incentives and, where the same company does eligible development work, alongside SR&ED. Capturing both means keeping the capital records and the R&D records aligned.

If you are planning equipment purchases in B.C., it is worth confirming eligibility before the purchase order is signed, not after.

Quebec's Development of E-Business credit, the CDAE, has long been one of the province's most valuable supports for technology companies: a credit of up to 30%, to a maximum of $25,000 per eligible employee each year, for qualifying software publishing, IT services, and e-commerce work carried out from a Quebec establishment.

The program is now evolving. Quebec is extending the credit to reward the adoption and integration of artificial intelligence under the CDAEIA stream, reflecting where the province wants its technology sector to invest. The mechanics follow the CDAE model: the credit is employee-based, activity-tested, and certified in advance.

That certification step is the part claimants underestimate. Before any credit can be claimed, the company must obtain an eligibility certificate from Investissement Quebec, covering both the corporation and the employees whose work qualifies. The activity tests are specific, the definitions are technical, and the application deserves the same care as the claim itself.

Program parameters continue to be adjusted as the CDAEIA transition proceeds, so current-year rates and thresholds should be confirmed at the time of filing. If your company builds software in Quebec and has never looked at the CDAE, or looked once and gave up, the AI expansion is a good reason to look again.

In a regulated environment, under Health Canada or GMP, you learn to document as you go, because you have to. Batch records, deviations, corrective actions, dated and signed. You cannot reconstruct them later. That same habit is the single best protection for an SR&ED claim.

When the CRA reviews a claim, it is not looking for a polished report written after the fact. It is looking for evidence that the work happened the way you described, when you said it did. Records made at the time carry far more weight than a story assembled at year-end.

You do not need a laboratory-grade system. A few simple habits cover most of it. Keep dated notes of what you were trying and why. Record the trials that did not work, not only the ones that did, because the failures are the proof of uncertainty. Keep versions of designs, formulations, or code so the changes are visible. And track, even roughly, who spent time on the work and when, so the labour side of the claim is supported.

The companies that find SR&ED easy and low-risk are usually the ones that built these small habits into how they already work. That is exactly what contemporaneous documentation support sets up: the evidence exists at the moment it is created, and each year's claim is prepared from real records rather than memory.

Formulation work is often treated as routine, and sometimes it is. Swapping one approved ingredient for an equivalent, or making a known adjustment, is not R&D. But a great deal of formulation work involves genuine uncertainty, and that part can be eligible.

Consider what makes formulation hard. Keeping an active ingredient stable over shelf life. Achieving a texture, taste, or delivery without a known recipe. Reformulating to remove an ingredient while holding performance. Meeting a regulatory limit that the existing formula cannot. In these cases the outcome is not certain at the start, and getting there takes a series of trials, measurements, and adjustments.

That is the line. If the result was predictable from what is already known, and you were applying a standard approach, it is routine. If the outcome was uncertain and you had to investigate systematically to reach it, it is likely experimental development, and likely eligible.

For natural health, cosmetic, food, and cannabis companies, this is one of the most commonly missed sources of SR&ED, precisely because formulation feels like everyday work. It is worth looking at each development project and asking a simple question: did we know this would work before we started? If the honest answer is no, there may be a claim in it.

A lot of the companies that qualify for SR&ED never claim it, because they do not think what they do counts as research. Manufacturers and natural health product companies are near the top of that list.

Here is the kind of work that often qualifies, even though it rarely gets called R&D. Reformulating a product to meet a new spec, a new regulation, or a new ingredient restriction. Scaling a process from a small batch to full production when it does not simply scale. Solving a stability, shelf-life, or consistency problem that standard methods did not fix. Reducing waste, energy, or cost in a process without losing quality. Hitting a tighter tolerance or a new performance target that took real trial and error.

In each of these, the same pattern appears: you set out to achieve something, the path was not obvious, and you had to experiment to get there. That is the heart of what SR&ED rewards.

The reason so much of this goes unclaimed is that it feels like ordinary problem-solving, not research. It is worth a second look. A short conversation, or the eligibility check on this site, will usually tell you quickly whether there is a claim worth preparing.

Ontario's Made Manufacturing Investment Tax Credit returns 10% of qualifying capital investment, refundable, up to $2 million a year. For a manufacturer investing in its plant, that is real money attached to spending that was going to happen anyway.

The credit covers two classes of investment: buildings (Class 1) used for manufacturing or processing in Ontario that qualify for the federal additional capital cost allowance, and machinery and equipment (Class 53, and its successor class for later acquisitions) acquired for use in manufacturing or processing of goods in the province.

Eligibility is deliberately broad: a Canadian-controlled private corporation, not exempt from Ontario corporate income tax, carrying on business through a permanent establishment in Ontario. The claim runs through the corporate tax return.

The planning point most companies miss is the interaction with SR&ED. The same plant that buys new equipment is often solving process problems that involve genuine technological uncertainty: scale-up that does not behave, tolerances that standard methods cannot hold, automation that has to be adapted rather than installed. The capital credit and the SR&ED credit reward different parts of the same investment, and a company doing both should be claiming both.

There is a common worry among business owners: "we are not doing science, so we cannot have SR&ED." It comes from thinking research only happens in a university lab. The program does not see it that way.

Academic research is usually driven by a question about how the world works, and the goal is knowledge that gets published. Industrial R&D is driven by a product or a process that has to work, on a deadline, at a cost. They feel like different worlds. But underneath, the thing SR&ED cares about is the same in both: was there technological uncertainty, and did you investigate it systematically?

SR&ED does not reward academic novelty. Your work does not have to be new to the world, and it does not have to result in a publication or a patent. It has to be work where the answer was not readily available to you, and where you worked through it in an organised way: forming an idea, testing it, and adjusting based on what you found.

That is why a machine shop solving a tolerance problem, a food company fixing a stability issue, or a natural health company reformulating to hold an active ingredient can all be doing SR&ED, even though none of them would call themselves scientists. The label does not matter. The uncertainty and the systematic work do.

If your team has ever said "we were not sure this would work, and it took us a few tries to get there," that sentence is often the start of an eligible project.

Many R&D tax claims are written by people who have never run an experiment. That sounds like a small thing, but it changes what gets captured and what gets missed.

In research, you are trained to see uncertainty clearly. You start from a question that existing knowledge cannot answer, form a hypothesis, design experiments, and expect some of them to fail. The failures are not a problem to hide; they are the evidence that the outcome was genuinely uncertain and that you worked systematically to resolve it. That instinct, to notice where knowledge runs out and to document the path through it, is exactly what a strong SR&ED claim is built on.

Industry R&D looks different on the surface but follows the same logic. On a production line, in a formulation lab, or in a software build, teams routinely run into problems where the standard approach does not work and the answer is not obvious. They try something, measure it, adjust, and try again. To the people doing it, this often feels like just getting the job done. To a trained researcher, it is recognisable as experimental development, and much of it is eligible.

This is where most claims are won or lost. The CRA's test is not about whether the work was new to the world, or whether it succeeded. It is about whether there was genuine technological uncertainty and a systematic effort to resolve it. Telling routine work apart from eligible investigation takes someone who has done the work, not someone matching activities to a checklist.

A researcher reads your project the way a reviewer will. Where was the uncertainty, and can we show it? What was the hypothesis, even if no one called it that at the time? Which trials and iterations are the evidence? Answering those questions honestly is what turns a pile of activity into a claim that holds up when the CRA asks.

Innovation rarely happens inside a single company. The strongest R&D programs we see draw on a wider ecosystem: a university group with the right instrument, a graduate researcher with the right training, a specialist consultant who has solved the problem before.

Grantwork's roots run through nearly a decade in academic research, and the practice maintains active connections with scientists, university research groups, and industry specialists across Canada, alongside a network of experienced consultants. For clients, that network shows up in practical ways.

It can mean structuring a research partnership through Mitacs, which funds internships that place graduate researchers inside companies to work on real technical problems, with the university sharing the cost. It can mean an introduction between a company and an academic lab whose capabilities fill a gap the company cannot justify building in-house. And it can mean bringing specialised scientific judgment into an SR&ED engagement when a project sits deep in a particular discipline.

Funding, in other words, is only part of an innovation strategy. Knowing who to work with is the other part, and it is one of the quiet advantages of working with an advisor who has lived on the research side of the ecosystem, not just the filing side.