Amazon’s rapid AI expansion is exposing growing cracks in its internal systems. An internal document seen by Business Insider highlights how the company’s push into artificial intelligence is driving widespread duplication of tools, scattered data, and overlapping platforms, a problem the report says is steadily getting “worse from both directions.”
An internal “Amazon confidential” report from February warns that the company’s AI push is accelerating tool duplication across teams, stating that “AI is making our tool duplication problem worse” as new systems are built faster than old ones are cleaned up.
At the core of the issue is speed, AI now lets engineers build software in minutes instead of weeks, encouraging teams to create new tools rather than reuse existing ones. Earlier, higher development costs naturally limited redundancy, but those constraints have now largely disappeared.
Amazon’s decentralized “two-pizza team” structure further amplifies the problem, making it difficult to track, coordinate, or consolidate thousands of independently built tools.
The report also points to a bigger data issue: AI tools can keep copies of information even after the original data is deleted or made private. In one case, Spec Studio reportedly showed private code that should have been removed. To fix this, Amazon is now looking at using more AI tools to spot duplicates, flag risks, and reduce overlap across its systems.