The verdict in three sentences
A recruitment firm holding 45,000 CVs yet still taking 15 days to present a shortlist does not have a talent pool problem, it has a pool exploitation problem. A custom ATS at 130,000 to 260,000 CAD indexes CVs by skill, surfaces the best profiles through AI-assisted search and shares annotated shortlists with clients. Built from day one around PIPEDA and Ontario privacy expectations, it becomes an asset of the firm rather than one more subscription.
SaaS or custom: what a 12-consultant firm really pays
SaaS ATS platforms charge between 120 and 230 CAD per consultant per month in 2026. For 12 consultants, that is 17,300 to 33,100 CAD a year, without ownership of structured data or control over the search logic. They suit a firm that is starting out. They become limiting when the talent pool is the firm's core value and consultants spend their days rewriting Boolean searches.
| Criterion (2026 estimate) | SaaS ATS | Custom ATS |
|---|---|---|
| Entry cost | Low, live in 2 weeks | 130,000 to 260,000 CAD |
| Recurring cost for 12 consultants | 120 to 230 CAD per consultant per month | Hosting and maintenance, priced in the quote |
| Pool search | Keywords and filters | AI-assisted skills search |
| Client shortlists | PDF export or basic link | Client portal with comments and decisions |
| Data hosting | Often outside Canada | Chosen to meet your residency needs |
| PIPEDA compliance | Check contract by contract | Built in: consent, retention, access requests |
| Ownership | Data exportable, logic not | Code, data and logic |
The features that save days
The time saving does not come from one feature. It comes from the chain. Incoming CVs are parsed automatically: skills, industries, years of experience, languages, mobility. Each candidate record is enriched with past interviews and client feedback. When a mandate arrives, the consultant describes the role in plain language and the tool proposes the 30 closest profiles, ranked and explained. The consultant refines, calls, then publishes a shortlist in the client's space, where the hiring manager comments and decides online.
| Mandate stage | Before (days) | With the custom ATS (days) |
|---|---|---|
| Initial pool search | 4 | 1 |
| Phone screening | 5 | 3 |
| Writing candidate summaries | 3 | 1 |
| Sending and client back-and-forth | 3 | 1 |
| Total time to shortlist | 15 | 6 |
Privacy compliance and timeline
In Canada, a candidate database is personal information under PIPEDA, and Ontario employers increasingly expect suppliers to align with provincial guidance, including disclosure of AI use in hiring under Ontario's 2024 Working for Workers rules that apply to job postings from 2026. The tool must handle candidate consent, a defined retention period (often 24 months after last contact), access and correction requests, and a log of every CV shared with a client. If part of the AI processing runs outside Canada, that transfer must be disclosed. These rules are designed upfront, they are hard to retrofit. In practice, budget 5 to 8 days of the project for privacy design: consent wording, retention jobs, access request workflow and a sharing log your clients can audit. Firms that skip this step usually pay two to three times more to fix it once the pool is live and clients start asking questions.
The project runs over 3 to 5 months: scoping and skills model (3 weeks), migration and parsing of the 45,000 CVs (4 to 6 weeks), search and client portal (6 to 8 weeks), pilot with 3 consultants then rollout. Our applied AI services page details the semantic search approach.
Mini case study
Emily, founder of an executive search firm in Toronto, runs 140 mandates a year with 12 consultants at an average fee of 28,000 CAD. Her consultants spend about 9 working days per mandate before the shortlist.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Calculation: cutting that to 4 days frees about 5 days per mandate, or 700 days a year across the firm. If only half is reinvested, that allows about 39 extra mandates (350 days divided by 9). At 28,000 CAD each, potential additional revenue exceeds 1,090,000 CAD. A 200,000 CAD project pays back in under 3 months of full activity.
FAQ
Can our 45,000 existing CVs be migrated?
Yes, migration covers text extraction, skills parsing and deduplication. Allow 4 to 6 weeks, with typically 5 to 10 % duplicates removed.
Does the AI pick candidates instead of the consultant?
No, it ranks and explains the 30 closest profiles, then the consultant decides. This keeps a human in the loop, limits bias and supports transparency obligations on AI use in hiring.
What do clients see in the shortlist portal?
A summary per candidate, the CV anonymised or not as you choose, and decision buttons. In practice, 70 to 80 % of client feedback arrives within 48 hours instead of a week by email.
How long should CVs be kept?
Set a retention period in your privacy policy, often 24 months after last contact. The tool requests consent renewal, then deletes or anonymises automatically.
What if we grow from 12 to 25 consultants?
A custom build is not billed per seat. Moving to 25 consultants mainly changes hosting, whereas a SaaS bill would roughly double.
Let's scope your project. We scope your skills model, talent pool migration and client portal, with a 3-consultant pilot and rollout in 3 to 5 months. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
