Dynamic and algorithmic pricing in Canada: what founders should know before letting AI set prices

Anyone who has refreshed a ticket page for a big tour has watched prices jump. Ticketing is the example most people know, but the same tools now set ride-share fares, hotel rates, e-commerce discounts and SaaS pricing experiments.

Three kinds of pricing

• Dynamic pricing changes prices in real time based on conditions such as demand, time of day or remaining inventory.
• Algorithmic pricing is broader: any system, including AI, that sets or recommends prices from data.
• Personalized pricing tailors the price to an individual based on information about them, such as browsing history, location or device. Critics call it surveillance pricing.

What the Competition Bureau is doing

The Bureau published a discussion paper on algorithmic pricing and competition on June 10, 2025 and took comments until August 4, 2025. On January 22, 2026 it published a What We Heard report summarizing 103 responses: 77 from individuals and 26 longer submissions from other stakeholders.

Individuals raised unfairness, privacy and lack of transparency. Other stakeholders pointed to efficiency benefits but also to tacit collusion when companies use similar data or the same third-party pricing service, and to personalized pricing without transparency, which some compared to drip pricing.

The legal risks

Prices that can’t actually be paid

The Competition Act (section 52(1.3) and section 74.01(1.1)) treats advertising a price that isn’t attainable because of mandatory added fees as false or misleading. That’s drip pricing, and the only exception is fees imposed by government, such as sales tax. An algorithm that shows one price on a listing and a higher one at checkout can create the same problem. Software-generated “limited time” or “only 2 left” messages are still your representations.

Collusion through shared tools

If several competitors feed data into the same pricing provider and their prices move together, regulators may ask whether the tool is doing the coordinating. Be careful with vendor contracts that pool competitor data, and get advice before joining any industry pricing platform.

Personal data

Personalized pricing runs on personal information, which brings in PIPEDA and, for Quebec customers, Quebec’s private sector law. Quebec requires you to tell a person when a decision about them is based exclusively on automated processing of their personal information (section 12.1). Your privacy policy should disclose personalized pricing, and you need a lawful basis for the data you use.

Interface design

Countdown timers, pre-checked add-ons and confusing price displays can cross into dark patterns. If they create a false or misleading impression, the Competition Act’s deceptive marketing rules can apply.

Before you let AI set prices

• List the data inputs your pricing tool uses and remove anything you can’t justify or disclose.
• Make sure the first price a customer sees includes all mandatory fees.
• Review vendor terms for data pooling with competitors, and keep a human able to override prices.
• Explain dynamic pricing in your terms of service and at checkout in plain language.
• Test prices across user groups to catch discriminatory outcomes before customers do.

Price testing in SaaS

You can test prices with care. Keep advertised prices accurate, honour the price a customer agreed to, and make sure your SaaS agreement explains how and when prices can change at renewal. For consumer plans, provincial subscription rules also apply. See the contract stack for software startups.