By mid-2027, fewer than 2 firms will publicly confirm fully dropping copilot-style tools — resolves 2027-07-14

The debate: whether companies will replace their coding assistants wholesale, moving from copilot-style tools that complete code inside the editor to fully autonomous agents.

Why it matters: whether real migration cases exist is the hardest evidence for telling coexistence from replacement, stronger than any benchmark score.

Who is on each side: the camp that expects agents to replace copilot-style tools, versus our call that the two will coexist.

Our call: by 2027-07-14 there will be fewer than two publicly verifiable, named cases of a company fully migrating (cancelling its copilot-style subscription and switching to a pure agent product). The company size threshold is 1,000 or more employees, or annual revenue of $100 million or more. Our confidence is 50%.

Sources still being linked up — the node(s) behind this call don't have a clickable original link resolved yet. Logged for follow-up.

Compile the migration events disclosed in public reporting, official case studies and customer financial filings, naming each case one by one: fewer than two counts as a hit; two or more counts as a miss. Noted plainly: we have not checked whether any such case already exists today. This is a forward count, so it needed no baseline figure when it was written.

What would overturn it

Two or more named cases would flip the reading that the two product types coexist without a migration wave.