Trane Technologies Achieves 60x Faster Building Insights with Amazon Bedrock AgentCore

Trane Technologies manages millions of connected heating, ventilation, and air conditioning (HVAC) assets worldwide. Previously, obtaining a single operational answer required cross-referencing multiple dashboards and drilling through menus for 20 minutes or more. This friction slowed operations and deferred corrective action, impacting the enterprise materially.

In 3–4 weeks, Trane’s engineering team built an AI-powered agentic solution on Amazon Bedrock AgentCore, reducing the 20-minute multi-screen diagnostic workflow to a 20-second natural language interaction. This represents a 60x improvement in time-to-insight, helping shift operations from reactive response to proactive, data-driven optimization.

The solution was built with the following architectural approach and key design decisions:

• Separating agent logic from tool execution.
• Integrating real-time telemetry through a centralized tool gateway.
• Tailoring responses to different personas.

The team phased the rollout, deploying internally first to capture critical technical feedback and refine accuracy before releasing externally to Trane Cloud customers. Role-based access and responses were implemented, with different roles receiving different tool and response formats. Observability was key, with end-to-end invocation tracing through CloudWatch enabling quick identification and resolution of issues.

Future phases of the solution will include:

• AgentCore Evaluations, a capability of Amazon Bedrock AgentCore, to gate rollout phases on automated accuracy thresholds.
• Policy in Amazon Bedrock AgentCore to replace custom authorization logic, making it easier to onboard new roles and audit access.
• Opening the AgentCore Gateway to other engineering teams across Trane, enabling them to build their own agents on the same centralized data layer.
• Adding new tools at a rapid pace, such as automating deep cost savings analyses for each customer site through natural language interaction.

This solution demonstrates the power of Amazon Bedrock AgentCore in accelerating business transformation and improving operational efficiency.

Source: aws-ml