A returning cook's food-testing claim gets picked apart by labour category

The situation

For an individual, the CRA gives you until the later of ninety days from the date on the notice and one year after the filing due date for the return in question to object, and if even that date passes, there is a further year in which to ask for an extension. Selam's filing-due-date window had already closed by the time the notice arrived, so the ninety-day clock was the one that governed her file, and by the time she opened the notice properly, a little over three weeks were left. She had moved back to Canada a few years earlier after time working abroad, settled in Whitby, and kept her job as a line cook while building a small side venture testing shelf-stable spice blends for a wholesale grocery buyer. The blends needed repeated batch testing to solve a spoilage problem, and under the research and development tax credit program, some of that testing work could be claimed as eligible labour.

She had filed the claim herself the previous year, working from templates she found online, and it had gone through without much scrutiny. This year the file landed with a reviewer named Dilshan, who pulled it for a closer look and wrote back that a portion of the claimed labour did not qualify as she had described it. The dollar amount was not large, under fifteen thousand dollars total, but Selam had already spent that money on ingredients and testing supplies and did not have it sitting in reserve to repay if the credit shrank.

The complication was who had done the work. Selam had paid Hanna, a friend who worked as a dental assistant, to help with recipe trials on weekends, treating the payments as contractor fees on the claim. Dilshan's letter suggested the working relationship looked more like employment than a contract for services, which changes how the wages get treated for the credit and for payroll obligations generally, and raised the possibility of a further review into whether payroll deductions should have been withheld at all.

Selam also worried, reading the letter a second time, about a line near the bottom noting that the reviewer might request additional documentation covering other parts of her return if the labour question was not resolved cleanly, language that read to her like a warning the file could widen rather than narrow.

Selam came to us with the letter and the deadline both in hand, unsure whether to simply accept the reduction or push back, and worried that disputing it might trigger the broader look at the rest of her tax filings that the letter seemed to threaten. She had never dealt with a tax dispute before, and the unfamiliar terminology in Dilshan's letter, phrases like eligible labour and arm's length arrangement, made the whole process feel harder to navigate than the actual dollar amount at stake probably warranted.

What the documents showed

We started by pulling together everything Selam had kept from the testing period: emails scheduling sessions with Hanna, a rough log of which batches were tested on which dates, and e-transfer records showing the payments. The picture that emerged was mixed. Hanna set her own hours around her regular dental assistant job, used her own kitchen for some of the trials, and was paid a flat amount per session rather than an hourly wage, all of which supported treating her as a contractor. But Selam had also directed exactly which formulations to test and in what order, supplied all the ingredients, and required Hanna to log results using Selam's own spreadsheet format, which cuts the other way.

Dilshan's position, laid out in the reassessment letter, was that the direction and control Selam exercised over the work outweighed the scheduling flexibility, and that Hanna's labour should be treated as employment income rather than contracted research labour. That distinction matters for the credit because eligible labour costs are calculated differently depending on whether the person doing the work is an employee of the claimant or an outside contractor, and the documentation standard expected of each is different too, with employment relationships typically requiring payroll records rather than informal invoices.

We also found a gap that worked against Selam: she had no written agreement with Hanna describing the arrangement, no invoices from Hanna for the sessions, and no record of Hanna having reported the income as self-employment on her own return. Those are the kinds of details a reviewer looks for to confirm a contractor relationship actually existed on paper, not just informally between two friends who happened to also be working together.

Reading through Dilshan's notes more closely, it became clear the reallocation was not based on any suspicion of fraud or exaggeration. The concern was narrowly about classification, not about whether the testing sessions had genuinely happened or whether the spoilage problem was real. That distinction mattered strategically, because it meant the argument we needed to make was about how to characterize an existing set of facts, not about defending facts that were themselves in question.

What helped was that Selam's underlying testing work was well documented and clearly experimental, with a real spoilage problem she was trying to solve through repeated trial, batch after batch, adjusting one variable at a time. The dispute was narrower than it first looked: it was about how one portion of the labour got characterized, not about whether the research itself qualified for the credit at all, which meant the bulk of the claim was never seriously at risk.

What we did

1. Confirmed the objection deadline and filed a protective notice first. With three weeks left on the ninety-day clock, our first move was to lock in Selam's right to dispute the reassessment rather than spend that time gathering documents before filing anything, since missing the deadline would have closed off every other option regardless of how strong the underlying facts eventually turned out to be.
2. Rebuilt a full timeline of the testing sessions from Selam's records. We cross-referenced her spreadsheet logs, the e-transfer payment dates, and Hanna's own calendar entries to produce a session-by-session account of what work happened when, which gave the file a level of detail it had not had when Selam filed the claim on her own the previous year.
3. Assessed the contractor-versus-employee question honestly before arguing it. Rather than assuming Selam's original characterization was correct, we weighed the control, scheduling, and equipment factors session by session against each other, so we knew going in which parts of Dilshan's position we could realistically contest and which we would likely have to concede.
4. Interviewed Hanna directly to fill gaps the paperwork could not. Because so much of the classification question turned on how independently Hanna had actually worked, we spoke with her separately to understand which sessions she had run on her own initiative and which ones Selam had closely directed, adding first-hand detail the spreadsheet logs alone could not provide.
5. Prepared a written agreement reflecting the actual arrangement going forward. Selam had never put anything in writing with Hanna, so we drafted a short agreement describing the contracted scope of future testing work, which did not change the past year's facts but showed Dilshan that Selam understood the distinction being drawn and had corrected course for the future.
6. Made the case for a split treatment rather than an all-or-nothing outcome. We argued that some sessions, where Hanna worked independently and invoiced informally, supported contractor treatment, while conceding that other sessions, run under closer direction with supplied ingredients, looked more like employment, aiming for a partial reallocation instead of losing the whole labour claim outright.
7. Corresponded directly with Dilshan through the objection period. Because the file had been pulled for review rather than moving straight to an automatic reassessment, there was room for genuine back-and-forth, and we used that window to answer follow-up questions promptly so the file did not stall waiting on our end.
8. Tracked the government's own processing timeline and kept Selam informed throughout. The review itself moved slowly once our objection was filed, and a large part of managing the file well was simply explaining to Selam, month after month, why nothing appeared to be happening while the queue worked through it in the background, and reassuring her that a long silence from Dilshan's office was ordinary at this stage rather than a bad sign about how the file was ultimately heading.

The outcome

The review concluded roughly eight months after we filed the objection, longer than Selam expected, mostly because of a backlog on Dilshan's end of the queue rather than anything specific to her file. The result was a split outcome rather than a full win: Dilshan accepted contractor treatment for the sessions where the documentation supported it, roughly two-thirds of the total labour claimed for Hanna's work, but held firm on the sessions run under closer direction, treating that remaining portion of the payments as employment income instead.

The credit Selam ultimately received was reduced from her original claim, and she owed a modest amount back along with a small amount of interest for the period the disputed portion had been outstanding since her original filing. It was not the full claim she had filed for, but it was a meaningfully better result than accepting Dilshan's initial position outright, which would have disallowed a much larger share of the labour cost and treated almost all of Hanna's work as employment income rather than contracted research.

Selam also avoided the broader review the second letter had hinted at. Because the labour question was resolved through documentation and cooperation rather than dispute, Dilshan closed the file without expanding into the rest of her return, which had been one of Selam's biggest fears going into the process.

Selam described the experience afterward as a hard but useful lesson in keeping proper records from the start. Going forward, she uses the written agreement template with anyone she brings on for testing work, and she keeps invoices and a clearer log rather than relying on informal arrangements between friends. The core research work she is doing continues to qualify for the credit, and the process this time cost her time, some interest, and a reduced credit rather than the loss of the claim entirely.

What you can learn from this