Microsoft's Bundle Strategy Spawns the Competitors It Defeats

Microsoft is consolidating Copilot into a single application launching this year, bundling chat, code generation. Autonomous agents under one roof.

This is presented as efficiency, as the natural evolution of AI. What it actually represents is a company repeating the exact strategic move that created both its greatest market dominance and its most persistent vulnerability—a lesson from 1990 that nobody seems to remember.

In 1990, Microsoft bundled Excel, Word, and Access into Office. Separately, these were good products competing against specialized competitors. Together, they became something else that locked customers in through purchasing convenience and switching costs. Within five years, Office had 90 percent market share in business productivity software.

The mechanism was pure leverage. They would make the integrated product so convenient and mutually reinforcing that leaving any single component means leaving them all. The lock-in is not malicious. It is structural. Once your data lives in Outlook, your spreadsheets in Excel, your documents in Word, the friction of migration becomes prohibitive—you stop evaluating whether each tool is best-in-class and evaluate whether the bundle is good enough.

But good-enough bundles spawn best-of-breed point solutions precisely because users eventually discover they are willing to pay extra to escape the bloat. Salesforce launched in 1999 because Excel-plus-Access was not the way to manage customer relationships. Slack launched in 2013 because email-plus-chat-in-Office was worse than a tool built for async communication alone. Users defected not because Office was bad. Because specialized alternatives were better, even at higher cost and new switching friction.

Now Microsoft is bundling again. Chat, coding, autonomous agents. One app. One login. One way to think about AI.