Database › Amazon
Amazon
A November 2025 internal memo made Kiro Amazon's recommended AI development tool and said Amazon does not plan to support additional third-party AI development tools. That is a tool-standardization policy, not a reported usage mandate. In 2026 Stores engineering set AI-native targets for teams, but a spokesperson said Amazon does not 'centrally mandate that teams use AI tools'.[2][3][4]
Tools
Metrics tracked
- Amazon estimated time saved from the number of Java dependencies migrated, and cost savings from hosts removed after the Java 17 upgrades.[1]
- A February 2026 internal document asked 2,100+ Stores engineering teams to triple release velocity, with at least 25 teams aiming for 10x, and set a goal of 80% of teams on AI-native practices (about 60% were at the time).[4]
- Tracked metrics include monthly active users, NPS, a 'Value Deriving Event' and deployments per engineer per week. The document flags Goodhart's Law as a risk.[4]
Guardrails
- The memo states: 'While we continue to support existing tools in use today, we do not plan to support additional third party, AI development tools.'[2]
- According to a memo reviewed by Reuters, Amazon revised its internal guidance for OpenAI's Codex to 'Do Not Use' in October 2025.[2]
- After retail outages, including one tied to Amazon Q, a 90-day policy for about 335 Tier-1 systems requires two-person review, documented approval through Modeled Change Management, and automated reliability checks. Amazon said only one reviewed incident was AI-related.[5]
- Amazon's six AI-native engineering tenets include 'No black boxes'.[4]
Hiring, reviews & process changes
- After employee pushback on top-down targets, the program shifted toward 'collaborative AI practices'.[4]
Reported results
- Amazon says Amazon Q Developer helped it migrate tens of thousands of production apps from Java 8/11 to Java 17. It puts the savings at more than 4,500 developer-years and $260M a year.[1]
- AI Teammate, a Slack-integrated agent, has expanded to 700+ active teams, and some groups, including parts of AWS, have adopted Pippin (ideas to technical designs) more broadly.[4]
Metric tags: Time saved, Cost savings, Adoption rate, PR throughput
Editor's note: The Reuters original returned HTTP 401 to automated checks. Use the syndicated CNA copy, which loads.
Sources
- AWS DevOps Blog: Amazon Q Developer just reached a $260 million dollar milestone · 2024-08-01 · company source
- Reuters via CNA: Amazon pushes in-house AI coding tool Kiro over competitors, memo shows · 2025-11-25 · press source
- Reuters (original; may be paywalled or block automated checks) · 2025-11-25 · press source
- Business Insider: Amazon pushes AI use and closely tracks adoption, as some employees push back · 2026-04-27 · press source
- Business Insider: Amazon orders 90-day reset after code mishaps cause millions of lost orders · 2026-03-10 · press source
Last verified 2026-10-25. Industry and size are editorial bands; see methodology. Spot an error? Submit a correction.