On 20 October 2026, Interim Commissioner of Competition Jeanne Pratt told the Canadian Chapter of the International Institute of Communications in Ottawa that "competition is not a destination, it is a process." The speech is mostly about telecoms. It looks back at the Bureau's opposition to the Rogers–Shaw merger and at switching rights in wireless and broadband. Its later sections, though, say where the Bureau thinks the next fights will be: data and artificial intelligence.
It names no technology platform. That restraint is informative, and it is the right place to start.
What the speech actually says
Under the heading "Data Portability and the Next Frontier of Competition," Pratt points to the Bureau's earlier report, Your Data, Your Control, which "examines how data portability can reduce switching costs and empower consumers." She compares it to telephone number portability, which "transformed what was once a significant switching barrier into a routine consumer expectation." She adds that portability works only when consumers trust it, which requires "strong privacy protections, robust security safeguards and interoperability across systems."
The next section, "AI, Innovation and Market Power," asks four questions. How might access to data affect competition? Will AI reinforce advantages held by large incumbents? Could algorithmic pricing systems create barriers to consumer information or switching? How do we preserve competitive markets while encouraging innovation? Pratt answers none of them. She says only that the challenge "is not to resist innovation" but to keep markets "open, contestable and competitive."
This is an agenda-setting speech, not an enforcement announcement. It proposes no new rule, case or legislative ask.
The strongest case for more intervention
The case for acting early is serious. Data advantages compound: firms with more users gather more data, train better models and attract more users. Switching costs are real, and they are hardest to see when they are informational rather than contractual. Algorithmic pricing can make it harder for a consumer to compare offers, even when no one has agreed to anything improper. Competition authorities that wait for conclusive harm are often late, because remedies applied after a market has tipped rarely restore the old rivalry. On this view, the Bureau is right to ask now.
Why the Bureau's own evidence points to a lighter tool
The Bureau's January 2026 study makes a more specific and more useful argument. It estimated that a data portability framework in insurance alone could save Canadians between $1.1 billion and $3.8 billion in annual costs. The estimate draws on behavioural economics research, a survey of more than 3,000 Canadians, and a comparison of regimes abroad. Lexpert's coverage notes that the EU, UK and Australia are already implementing similar frameworks.
Portability is the pro-innovation answer to the concerns Pratt raises. It does not penalise scale, restrict how firms build AI systems, or require regulators to second-guess product design. It lowers the cost of leaving, so a challenger can win customers on merit. That is how number portability helped wireless competition. It is also a remedy where the evidence is thin: if a market is not actually tipping, portability costs little. If it is, portability is the least intrusive way to keep it open.
The algorithmic pricing question deserves more caution. Pricing software is common, and most uses are benign: they help firms respond to demand and cut costs. The Bureau should separate three cases:
- Software that helps a firm set its own price, which is ordinary business.
- Software that conceals pricing information or makes comparison harder, which is a transparency problem.
- Software that coordinates rivals' prices, which falls under existing conspiracy and abuse provisions.
Only the second and third need attention. Treating all three alike would impose real costs on small firms that use off-the-shelf tools.
Trust is the binding constraint, and it cuts both ways
Pratt's remark that portability works only if consumers trust it is the most important line in the speech. Mandated data movement creates new security and privacy exposure. Canadians' willingness to use it will depend on whether the surrounding framework respects privacy.
That is awkward given other federal proposals. The Electronic Frontier Foundation has warned that Bill C-22, the Lawful Access Act, which has passed the House of Commons and is before the Senate, would allow government-ordered backdoors and require up to a year of metadata retention. That is the EFF's reading of a bill still moving through Parliament, and the government may amend it. Still, if data-sharing frameworks are to earn consumer trust, Ottawa cannot promote portability on competition grounds while weakening confidence in how data is held on other grounds. Competition policy and privacy policy have to be designed together.
What to watch
Three tests will show whether the speech becomes policy:
- A concrete portability proposal. The January report is a roadmap, not a statute. The meaningful signal is a sector-specific framework with named standards, a named supervisor and clear privacy rules.
- Evidence on AI market structure. The Bureau's question about incumbents is empirical. Canadian data on compute, data access and switching in AI services would help more than analogy to other markets.
- Proportionality in enforcement. Existing tools should be used first. New powers should wait until the Bureau shows a gap that portability and current law cannot close.
Pratt's framing, that the aim is to keep markets open rather than resist innovation, is the right one. The test is whether the Bureau's follow-through is as restrained as its language. Portability, transparency and interoperability can lower barriers without picking winners. Heavier ex ante rules on AI developers, adopted before any harm is demonstrated, would risk deterring the investment Canada wants. The speech leaves room for the first path, and the Bureau's own evidence supports it.