IT / Software
Example Vendor Inc.
- Concentration
- 9.2
- Behaviour
- 8.4
- Adverse
- 7.1
Captures 84% of one ministry’s sole-source IT spend. Three contracts amended above original bid in 2024. Two adverse-media hits this quarter.
CANveat puts a caveat on every Canadian government contract — surfacing concentration, amendment creep, and adverse media before a vendor becomes a scandal.
IT / Software
Captures 84% of one ministry’s sole-source IT spend. Three contracts amended above original bid in 2024. Two adverse-media hits this quarter.
The pattern is recognizable
ArriveCAN, McKinsey, GC Strategies, WE Charity. The data to spot the pattern is already public — federal Proactive Disclosure of Contracts, Alberta sole-source registries, the Integrity Regime ineligible-supplier list, the Lobbying Registry, news media. It just sits scattered across portals, in different schemas, with no canonical vendor identity, and nobody is reading it at decision speed.
How a $0.25M contract becomes a $59M scandal.
ArriveCAN, 2020–2024. Each step is an amendment to the same sole-source contract; each dot is a public adverse-media report; the right edge marks the day federal procurement was suspended.
How CANveat works
Canonical vendor identity across CRA, federal, and provincial datasets. The same vendor under five name variants in five tables resolves to one row, one score.
Composite 0–30 risk across concentration, behaviour, and adverse signals, calibrated against the patterns that surfaced in past procurement failures.
Briefing-style UX surfaces one recommended action per flagged vendor and produces a one-click memo formatted to match TBS conventions.

Use cases
Built for
About CANveat
CANveat was built for Agency 2026 — Ottawa, the Government of Alberta’s national AI-and-accountability hackathon (April 2026). The pan-Canadian framing leads with Alberta procurement data because it is the only true procurement-shaped open dataset in the provided platform, and the hosting context makes the alignment intentional.
Built by the CANveat hackathon team — names to be added.
Open-data attribution
dataset-analysis-platform/.No login. The data is open; the briefing is one click away.