How Miro uses Amazon Bedrock to boost software bug routing accuracy and improve time-to-resolution from days to hours

2026-05-11 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, long

Summary

Miro, an AI-powered innovation workspace serving over 95 million users, partnered with AWS to develop BugManager, an AI-powered solution for automated bug triaging. This system addresses the challenge of accurately routing software bugs to approximately 100 engineering teams, a process that previously led to significant delays and an estimated 42 years of lost productivity annually due to misrouting. BugManager utilizes Amazon Bedrock, specifically Anthropic's Claude Sonnet 4, and Amazon Nova Pro for multimodal parsing, combined with Retrieval Augmented Generation (RAG) to enrich bug reports with context from various knowledge bases like GitHub, Confluence, and Jira. This approach has resulted in six times fewer team reassignments and a five times shorter time-to-resolution, transforming bug resolution from days to hours, with a top-1 accuracy of over 75% and top-3 accuracy of 95%.

Key takeaway

For MLOps Engineers or AI Engineers tasked with improving internal support workflows, consider adopting an LLM-powered RAG system like BugManager. Your team can leverage Amazon Bedrock and its integrated services to build a zero-training, adaptable solution that significantly reduces bug misrouting and accelerates time-to-resolution, directly impacting developer productivity and product quality. Focus on robust context retrieval and transparent decision rationales to drive adoption.

Key insights

LLM-powered RAG systems can significantly improve bug routing accuracy and resolution times in complex software organizations.

Principles

• RAG enhances LLM performance by providing dynamic, up-to-date context.
• Multimodal parsing unifies diverse bug report data into a single modality.
• Transparent routing decisions increase user acceptance and trust.

Method

BugManager parses multimodal bug data, enriches it with RAG-retrieved context from internal knowledge bases, and uses an LLM (Claude Sonnet 4) with an optimized prompt for team classification and optional root cause analysis.

In practice

• Integrate Slack for frictionless user interaction with bug triaging.
• Use Amazon Bedrock Knowledge Bases for managed RAG workflows.
• Implement human-in-the-loop for higher accuracy with top-N predictions.

Topics

• Amazon Bedrock
• Bug Triaging
• Retrieval-Augmented Generation
• Anthropic Claude Sonnet 4
• Software Development Workflow

Best for: AI Engineer, Machine Learning Engineer, MLOps Engineer

Related on AIssential

• Transform Enterprise Search with GenAI: Build a Production-Ready RAG Knowledge Assistant Using… — RAG with Amazon Bedrock enhances enterprise search by providing accurate, grounded, natural language access to proprietary data.
• Amazon Bedrock AgentCore harness is now generally available: Go from idea to production-grade agent in minutes — Amazon Bedrock AgentCore harness abstracts complex infrastructure, enabling rapid deployment and management of production-grade LLM agents.
• LLM‑D Explained: Building Next‑Gen AI with LLMs, RAG & Kubernetes — LLM-D optimizes LLM inference by intelligently routing requests and disaggregating prefill/decode phases on Kubernetes.
• Agentic RAG Failure Modes: Retrieval Thrash, Tool Storms, and Context Bloat (and How to Spot Them Early) — Agentic RAG's control loop introduces fragility, leading to predictable failure modes without proper budgeting and stopping rules.
• Build an AI-powered website assistant with Amazon Bedrock — An AI-powered RAG system on AWS Bedrock can provide instant, role-based support from diverse data sources.
• Grounding LLMs with Fresh Web Data to Reduce Hallucinations — Grounding LLMs with live web data via managed search infrastructure significantly reduces hallucinations from outdated information.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.