Andrej Sokolow/picture alliance via Getty Images
- AI is helping employees create new apps and tools faster.
- That creative explosion comes with downsides. Some are calling this "AI sprawl."
- Amazon has seen a surge in duplicate internal tools and data. Its solution: More AI.
Amazon's AI boom is creating a new kind of mess: a growing bloat of internal tools and duplicated data.
Some teams are rapidly building their own AI-powered applications to automate workflows and organize information. But that creative explosion is also causing problems, such as software and data duplication, according to an internal document obtained by Business Insider.
"AI is making our tool duplication problem worse," the document stated. "More duplication is being created faster, and less of it is being cleaned up."
The trend points to a broader shift across corporate America. Generative AI is driving what some call "AI sprawl," a surge of AI tools and autonomous agents that risks overwhelming companies' centralized oversight and security controls.
As more employees spin up tools on their own, sometimes in minutes with AI assistants, organizations can lose visibility into which systems are in use, where sensitive data lives, and how much redundant software they maintain.
Cloud —> SaaS —> AI Sprawl
It's a familiar story with a twist. When cloud computing emerged 20 years ago, rogue employees fired up Amazon Web Services accounts without permission. Then, "SaaS sprawl" spread cloud software across companies with little oversight at first.
In both cases, companies eventually responded by reigning in extremes, building oversight, and designing new systems to take official advantage of these powerful new technologies. Generative AI is in the early stages of the same pattern, but it's moving a lot quicker.
The effects of this unruly, creative AI boom may be especially pronounced at Amazon. The company has embraced AI across its business, with CEO Andy Jassy urging employees to adopt the technology or risk falling behind.
'Dramatically lowers the barrier'
Amazon has, for years, grappled with teams building similar tools in parallel, according to the document obtained Business Insider. Those overlaps were costly but manageable. Building software required significant time and engineering resources, and maintaining it created enough burden that some redundant tools were eventually retired.
AI is changing that equation.
According to the document, AI "dramatically lowers the barrier to building new tools," allowing teams to prototype and ship software far more quickly. Instead of searching for existing solutions, engineers can create their own in a fraction of the time and maintain them at a much lower cost.
The result is a surge in overlapping systems, with less pressure to consolidate them. "AI is now making this problem worse from both directions," the document stated.
The document, marked "Amazon confidential," was produced in February by a team responsible for evaluating and improving AI tools used across Amazon's vast retail business, which spans thousands of engineers. The company encourages open dialogue about challenges, an approach that has helped it get ahead of problems in the past.
In an email to Business Insider, Amazon spokesperson Montana MacLachlan said the document reflects the perspective of a single team and that it is "inaccurate" to use one group's view to characterize the experience of the company's broader workforce.
'Artifacts persist'
The problem goes beyond tool overload. It's also introducing new risks in how data is handled and stored.
Many of Amazon's AI systems take in internal data and convert it to new formats, such as knowledge bases and summaries. Those outputs are often stored separately from the original source, effectively creating new copies of the same information, according to the document. If the original data is later deleted or access is restricted, those derived versions don't always update.
In one internal case, a system called Spec Studio continued showing software details that had been made private in Amazon's internal code repository, the document said. The company is now pushing teams to better document how they handle permission changes and data deletion.
"Any system that ingests data, transforms it through AI, and stores the output separately faces the same problem: when source permissions change or data is deleted, derived artifacts persist," the document stated.
Using AI to fix an AI problem
The rapid spread of AI-generated tools is creating "shadow AI" within organizations, where unauthorized applications introduce risks such as sensitive data exposure and regulatory violations, Debo Dutta, chief AI officer of cloud company Nutanix, told Business Insider.
"If not governed properly, this can all lead to data and system disruption," Dutta said.
Amazon's answer may be more AI.
The company is exploring ways to use AI to identify duplicate tools, flag risks, and nudge teams to consolidate earlier, before overlap becomes hard to unwind, according to the document.
The challenge is balancing speed with coordination. Amazon has long prized autonomy, with small, independent teams moving quickly and making their own decisions, through an approach often described as the "two-pizza team" model. That culture has fueled rapid experimentation, but it may also make the problem harder to solve.
"Teams that build bespoke AI systems are likely to repeat these same issues," the document said.
Have a tip? Contact this reporter via email at ekim@businessinsider.com or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address, a nonwork WiFi network, and a nonwork device; here's our guide to sharing information securely.