AWS Partner Network (APN) Blog
Category: Artificial Intelligence
Predictive Maintenance at Scale with Treon Flow and AWS IoT Core
Treon Flow extends Amazon Monitron sensor technology into a predictive maintenance solution on AWS, combining asset monitoring with machine learning analytics. In this post, we explore where industrial predictive maintenance is heading, from contextual analytics to automated maintenance workflows and technician productivity tools. We also look at how teams already using Amazon Monitron can extend their condition monitoring programs on AWS while keeping a familiar sensor design and ways of working.
Bringing AI code governance into Kiro with Qodo
AI coding agents now write code faster than pull request reviews can keep up. The Qodo Kiro power brings independent, context-aware code review into Kiro, so developers can check agent-generated code against their organization’s standards and wider codebase before a pull request opens. This post explains how Kiro powers work, how Qodo’s review loop runs inside the IDE, and how Qodo runs its inference on Amazon Bedrock. Teams keep one consistent quality bar across every change, whichever agent wrote it.
Reach more customers, close deals faster with new AI-powered capabilities in AWS Partner Central
New AI-powered capabilities in AWS Partner Central and pre-built Skills in AWS Quick help you find the right customers, launch campaigns in minutes, and accelerate co-sell with AWS. With Partner Central Agents working alongside you, you spend less time on manual work and more time growing your business, right in the tools you already use.
Precision Retrieval for Regulated Industries with Vespa.ai on AWS
Enterprise AI succeeds or fails on whether the right context reaches the model. This post looks at two regulated domains, life sciences and financial services, where an irrelevant answer slows a clinical trial or undermines an analyst’s conclusion. It shows why the winning pattern on AWS is use-case led: retrieve with hybrid signals, rank with domain logic, and serve in real time. See how Reveal HealthTech and RavenPack build precision retrieval and RAG with Vespa.ai.
Governing Agent Loops on Amazon Bedrock
Agent loops can automate real multi-step work, but running them in production raises questions about cost, actions, and audit. This post shows how to build and govern agent loops with Amazon Bedrock and AWS-native controls. It also explains how to recognize when governance is better handled by a dedicated control plane, such as Blunom, that extends those controls across teams and clients.
Runtime governance for AI agents with Noma Security and AgentCore Gateway
An AI agent can be authorized to use a tool and still use it for the wrong task. Noma Security’s AI Detection and Response (AI-DR) complements AWS controls with semantic evaluation of content and intent at Amazon Bedrock AgentCore Gateway. By connecting proposed actions with the user’s request, it helps catch intent mismatches such as a payment change during a summary task. Security teams gain configurable enforcement and a shared session history for supported traffic routed through the gateway.
Detect secrets before they ship with the GitGuardian power for Amazon Kiro
AI coding agents can write and commit code faster than pipeline scanners can check it, and hardcoded secrets are one of the first risks to surface. In this post, we show how the GitGuardian Kiro power detects secrets at the moment code is written, and how AWS Secrets Manager completes the remediation loop by storing, encrypting, and rotating replacement credentials, so exposed secrets become a code change, not an incident.
Running SnapLogic Jean-Paul in Production on Amazon Bedrock
Enterprises can build an AI agent that reasons impressively in a pilot, but moving it into production is where projects stall. The blocker is rarely the model, it is the infrastructure, security, and governance. This post shows how SnapLogic’s Jean-Paul runs inside your own AWS account on Amazon Bedrock, the production-readiness pattern for taking an enterprise AI agent from pilot to production on AWS.
Protecting Apache Iceberg Lakehouses at Scale with Rubrik on AWS
Apache Iceberg is now a standard open table format for AWS lakehouses, and because that data feeds AI pipelines and financial reporting, it needs stronger protection. Recovering a working table takes more than restoring files, because it depends on consistent metadata, manifests and catalog. Rubrik’s new Apache Iceberg Data Protection on AWS backs up the full table to the customer’s AWS account or Rubrik Cloud Vault, then updates the catalog on recovery so the table is immediately queryable. The post explains why native time travel alone falls short and how Rubrik and AWS bring enterprise cyber resilience to the lakehouse.
JA Canada launches generative AI with Qalius and Amazon Bedrock
JA Canada worked with Qalius Consulting Inc., a Canadian AWS Advanced Tier Services Partner, to move a generative AI project from design to deployment in just four months. The project focused on one clear use case: helping visitors find information across JA Canada resources through a conversational experience. This post shares four key ingredients for generative AI success: bounding the first use case, building in evaluation, treating safeguards and cost controls as launch requirements, and using managed AWS capabilities to reduce unnecessary engineering work.









