Seasonal Hiring in Burnaby: Screen and Onboard With AI

Holiday postings in Burnaby go up in late September. The math is rough. You need twenty five people in six weeks, you get four hundred applications, and the person reading them is also running the floor on a Saturday.

AI helps here. It also creates real exposure if you aim it at the wrong decision. Below is what the primary sources actually say, including the parts that argue against automating.

The labour market you're hiring into

Statistics Canada reported that job vacancies held steady at 510,200 in the second quarter of 2026, and that the average offered hourly wage for vacant positions rose 2.0% (up $0.55) to $28.55 (Statistics Canada, The Daily, September 15, 2026). B.C.'s general minimum wage went to $18.25 per hour on June 1, 2026 (Province of British Columbia).

Read those two numbers together. The average vacancy in Canada is now offering well above B.C.'s floor, so a seasonal role posted at minimum wage is competing from behind. Your edge is speed of response, not cleverness. That happens to be exactly what AI is good at, and it doesn't require the AI to judge anybody.

Five jobs worth handing to AI, and one to keep

• Intake and normalization. Pull availability, start date, transit access, certifications and past roles out of messy resumes into one structured table.
• First contact. Acknowledge every application within minutes, with the shift pattern, pay rate and location stated plainly, so mismatches opt out on their own.
• Scheduling. Run the back and forth on interview slots, send reminders, rebook the no shows.
• Onboarding packets. Generate role-specific orientation material and a per hire checklist.
• The FAQ. Answer the same twelve questions about parking, uniforms, payroll and the first shift, around the clock.
• Keep for a human: the actual yes or no on a candidate. Here's why.

Screening is where the evidence gets uncomfortable

Wilson and Caliskan tested embedding based resume screening across nine occupations using over 500 resumes and job descriptions, and found the models "significantly favoring White-associated names in 85.1% of cases and female-associated names in only 11.1% of cases," with Black male candidates disadvantaged in up to 100% of cases (Wilson and Caliskan, AAAI/ACM AIES 2024).

A newer audit points the other way. Gao, Jiang and Yan ran a paired resume design across 14 mainstream large language models at 24,024 paired postings per model, and found the single 2023 vintage model produced a pro-White callback gap of +2.12 percentage points, while every model released in 2024 or later showed either no significant gap or a reversal favouring Black applicants, reaching -3.01 percentage points (Gao, Jiang and Yan, arXiv, June 2026).

Those two results disagree, and the disagreement is the finding. Bias direction is not stable across model generations, retrieval versus generation setups, or prompt wording. If you can't audit your own pipeline, you can't claim it's fair on the grounds that a recent paper said the newer models were fine.

The workable compromise: let AI extract and organize facts, and let a person rank. "Holds FoodSafe, available Dec 20 to Jan 3, lives in New Westminster" is a fact. "Good culture fit" is a judgment, and it should have a human name attached to it.

The overnight batch is the cheap win

Processing a day's applications doesn't need instant answers. Anthropic's Message Batches API documents asynchronous processing "with most batches finishing in less than 1 hour while reducing costs by 50%", limited to 100,000 requests or 256 MB per batch, with a 24 hour expiry and results available for 29 days (Anthropic, Batch processing documentation). Other providers offer similar asynchronous tiers.

So queue the day's applications at 7pm and have structured summaries waiting in the morning at half the per token cost. Any dollar figure attached to this moves with every model release, so price it against the vendor's current page the week you build, not from a blog post.

Onboarding has a legal deadline in B.C., and AI can help you hit it

This is the part seasonal employers miss most. Under B.C.'s Occupational Health and Safety Regulation, an employer must ensure that before a young or new worker begins work, that worker gets health and safety orientation and training specific to that workplace. "Young worker" means anyone under 25 years of age, and "new worker" covers anyone new to the workplace (WorkSafeBC, OHS Regulation Part 3, sections 3.22 to 3.25).

Section 3.23 sets out what that orientation has to cover, including the supervisor's name and contact information, rights and responsibilities under workers' compensation law, workplace health and safety rules, hazards such as violence and working alone, personal protective equipment, first aid and injury reporting, emergency procedures, WHMIS information, and instruction and demonstration of the worker's own tasks. Section 3.25 requires the employer to keep records of all orientation and training provided.

A seasonal crew is, almost by definition, made up of young and new workers. AI is genuinely good at the drafting and the tracking: build the orientation content per role, push it before day one, chase the ones who haven't finished, hold the record. It can't do the hands on demonstration, and it can't be the supervisor whose name goes on the form.

The privacy rules that actually bind you here

B.C.'s Personal Information Protection Act defines employee personal information as information collected, used or disclosed solely for purposes reasonably required to establish, manage or terminate an employment relationship. Section 13 lets you collect it without consent where that collection is reasonable for those purposes, but you must notify the individual before you collect it. Section 35 says that where personal information has been used to make a decision that directly affects someone, you keep it at least one year so they have a reasonable chance to access it (Personal Information Protection Act, BC Laws).

Federal, provincial and territorial privacy regulators published joint principles for generative AI on December 7, 2023. The line that matters most in hiring: "accountability for decisions rests with the organization, and not with any ... automated system," alongside an expectation of human review or reconsideration of significant decisions (Office of the Privacy Commissioner of Canada and counterparts, 2023).

Ontario went further. Since January 1, 2026, employers there with 25 or more employees must state in a publicly advertised job posting that AI is used, if they use it to screen, assess or select applicants, and must keep postings and application forms for three years after public access ends (Government of Ontario, Your guide to the Employment Standards Act). That rule does not apply in B.C. It does apply if you post into Ontario, and it's a fair guess at where the rest of the country drifts. One honest disclosure line costs you nothing.

Where this doesn't apply

• Small volumes. Under roughly fifteen seasonal hires, setup time exceeds the savings. A shared inbox and a phone still win.
• When the offer is the problem. If applicants ghost because the shifts are bad or the pay sits at the floor, faster screening just sorts a shrinking pile more efficiently.
• Judgment heavy roles. Assistant manager, anyone handling cash alone, anything carrying a customer safety duty. Screen those yourself.
• When you can't explain the output. If you can't tell a rejected applicant what mattered, you're carrying the accountability the regulators place on you with no way to discharge it.
• Anything unaudited. Given how sharply the screening research disagrees with itself, pointing a model at candidate names and calling it neutral is not a defensible position.

One illustrative figure, not a promise: a shop that moves 6 hours a week of resume triage and scheduling email off a manager's desk across a 6 week ramp gets back about 36 hours. Your numbers will be different. Measure yours before and after rather than trusting anyone's estimate, including this one.

Sources

• Statistics Canada. "Job vacancies, second quarter 2026." The Daily, September 15, 2026. https://www150.statcan.gc.ca/n1/daily-quotidien/260915/dq260915a-eng.htm
• Province of British Columbia. "Minimum wage." Employment Standards, updated June 1, 2026. https://www2.gov.bc.ca/gov/content/employment-business/employment-standards-advice/employment-standards/wages/minimum-wage
• Wilson, K. and Caliskan, A. "Gender, Race, and Intersectional Bias in Resume Screening via Language Model Retrieval." Proceedings of the 2024 AAAI/ACM Conference on AI, Ethics, and Society, 2024. https://arxiv.org/abs/2407.20371
• Gao, Z., Jiang, W. and Yan, Y. "Can LLMs Hire Fairly? Racial Bias in Resume Screening." arXiv preprint, June 2026. https://arxiv.org/abs/2606.28978
• Anthropic. "Batch processing." Claude platform documentation, accessed September 2026. https://platform.claude.com/docs/en/docs/build-with-claude/batch-processing
• WorkSafeBC. Occupational Health and Safety Regulation, Part 3, sections 3.22 to 3.25 (Young and New Workers). https://www.worksafebc.com/en/law-policy/occupational-health-safety/searchable-ohs-regulation/ohs-regulation/part-03-rights-and-responsibilities
• Government of British Columbia. Personal Information Protection Act, SBC 2003, c. 63. BC Laws. https://www.bclaws.gov.bc.ca/civix/document/id/complete/statreg/03063_01
• Office of the Privacy Commissioner of Canada and provincial and territorial counterparts. "Principles for responsible, trustworthy and privacy-protective generative AI technologies." December 7, 2023. https://www.priv.gc.ca/en/privacy-topics/technology/artificial-intelligence/gd_principles_ai/
• Government of Ontario. "Requirements related to publicly advertised job postings." Your guide to the Employment Standards Act, in force January 1, 2026. https://www.ontario.ca/document/your-guide-employment-standards-act-0/requirements-related-publicly-advertised-job

If you're hiring across Metro Vancouver this season and want a second set of eyes on the pipeline before the rush hits, book a free call with Autana Solutions. We'll walk your current intake, screening, scheduling and orientation steps, tell you plainly which ones are worth automating this year and which ones aren't, and leave you with the plan either way.