A 40% discount that expires June 30. A price increase that kicks in July 1. A new $99/month bundle launching May 1 that bundles Copilot with enough other products to make the standalone price look reasonable by comparison. Microsoft has constructed a pricing architecture around Copilot that is, at its most charitable reading, a sophisticated urgency machine. The question enterprise buyers should be asking is not whether Copilot is worth $30 a month. The question is why Microsoft needs you to decide before summer.
Gartner's finding that 49% of AI programs stall when value is unclear is the most important number in this conversation, and Microsoft's promotional calendar is a direct response to it. If Copilot's productivity gains were self-evident, the company would not need hard deadlines to accelerate commitment. Discounts expiring June 30, base pricing rising July 1, a new premium tier arriving May 1: this is not a product launch cadence. It is a conversion funnel designed to force capital decisions before the ROI case has been made.
The Lock-In Is the Product
Microsoft spent roughly $80 billion on AI capital expenditure in fiscal 2025. That number does not get justified by $30/month subscriptions at current adoption rates. It gets justified by installed-base scale, by workflow dependency, by the organizational cost of switching once Copilot is embedded in how a company's knowledge workers operate. The discounts are not generosity. They are the cost of acquiring the dependency.
Consider what a 1,000-person enterprise actually commits to at the 30% discount tier: roughly $252,000 annually, locked in before most organizations have a rigorous framework for measuring whether AI assistance translates into billable hours saved, errors reduced, or decisions improved. Microsoft's own consumer terms of service describe Copilot as being for "entertainment purposes only," which is a legal hedge, not a product description, but it does raise a fair question about the evidentiary basis for the productivity claims being made to enterprise procurement teams.
To be fair: some organizations have documented genuine efficiency gains from Copilot, particularly in summarization-heavy workflows like legal review and customer support triage. The technology is not a fraud. But "works for some use cases under favorable conditions" is a different claim than "worth a six-figure annual commitment before you've measured your own outcomes."
What the Dragon Copilot Pivot Reveals
Microsoft's decision to simplify Dragon Copilot's healthcare pricing from per-action fees to per-encounter billing is a small but telling signal. The company recognized that granular, complex pricing creates adoption friction. The fix was structural simplification. That same logic has not yet been applied to standard Copilot licensing, where the discount tiers, bundle configurations, and seat minimums require a procurement specialist just to evaluate. Complexity that benefits the seller is rarely accidental.
Enterprise IT leaders who are currently evaluating Copilot should do one thing before signing anything: define the specific productivity metric they will use to evaluate the investment at 12 months, write it into the contract, and ask Microsoft to stand behind it. Not a vague commitment to "AI-powered productivity," but a number. Hours per week recovered per knowledge worker. Ticket resolution time. Something measurable. If Microsoft's sales team cannot engage with that conversation, the discount window is not the problem. The product's readiness for your organization is.
Microsoft is betting that organizations will rationalize the cost after they are dependent on the workflow. That bet has worked before, with Office, with Azure, with Teams. The question is whether you want to be the organization that proves it works again, or the one that waited for the evidence first.
