On 11 November 2026, AWS announced a new catalog of 23 production‑ready, open‑source reusable assets for generative AI development. The assets are housed in the AWS Generative AI Innovation Center, which has been incorporated into AWS Frontier AI Engineering & Services. The catalog is intended to allow teams to build on validated, security‑reviewed components rather than recreating common AI building blocks from scratch.
Asset Types and Validation
The catalog includes Retrieval Augmented Generation pipelines, document processors, agent scaffolding, evaluation frameworks and prompt‑management patterns. Each asset has undergone pre‑deployment reviews and ongoing maintenance to ensure quality, security and production readiness. Deployment guides are provided, and the assets are instrumented for adoption tracking, with comprehensive documentation and customer success stories.
Customer Use Cases
DoorDash deployed the Contact Center GenAI Agent to build a voice‑operated support solution that handles over 100,000 calls daily. The system achieves a 2.5‑second response latency and a 50‑fold increase in testing capacity. Lyft used the same agent to enable agentic AI for riders and drivers, demonstrating the asset’s versatility across customer engagement models.
Ricoh implemented the GenAI Intelligent Document Processing asset to automate document handling. The solution increased throughput sevenfold, from 10,000 to 70,000 documents per month, and reduced engineering hours per deployment from 80 to under 5 hours. Onboarding time dropped from 4–6 weeks to 2–3 days, saving 1,900 annual person‑hours.
Other assets in the catalog include the Fullstack AgentCore Solution Template (FAST), which enables rapid deployment of a secured React front‑end connected to Amazon Bedrock AgentCore in under 30 minutes using infrastructure‑as‑code, and the Agentic Chatbot Accelerator, which uses AWS CDK to deploy agentic chatbots with pre‑configured patterns for authentication, hosting and scaling.
The Contact Center GenAI Agent incorporates security controls via Amazon Bedrock Guardrails to detect hallucinations, filter content and enforce topic boundaries in real time.