ShipStation's Slack IT agent on Bedrock AgentCore resolves 97.5% of queries
AWS ML Blog · rss · 2026-10-13
An AWS ML Blog deep dive explains how ShipStation Global built Deckhand, an AI IT help desk agent for Slack on Amazon Bedrock AgentCore, resolving 97.5% of channel queries with 96% positive feedback over three months.
Context: the help channel saw 1,500+ threads in 12 months, with over a fifth arriving outside US business hours and answers scattered across Confluence, intranet pages, and old threads.
Architecture: a FastAPI/Slack Bolt service on App Runner handles events and user context; a lightweight Strands Agents SDK agent runs in isolated AgentCore runtime containers using Claude on Bedrock; a single Lambda exposed as MCP tools via AgentCore Gateway handles classification, Jira tickets, and image analysis; a Bedrock knowledge base indexes Confluence, intranet, and resolved Slack threads.
Notable choices: model ID via environment variable for easy swapping; every request carries the requester's Slack profile so tickets auto-fill; unanswered queries become fully contextualized Jira tickets in one click.
More from coding & agent
- FinOps for Agentic AI: taming skyrocketing cloud and token costs — DavidLinthicum · 2026-10-14
- QGIS bans all AI-generated code, and longtime users are walking away — MaxLenormand · 2026-10-14
- Gemini's Argon wrote 5,000 lines of Python to build its own ray-tracing engine — fofrAI · 2026-10-14
- Developer's AI agents built, critiqued and rebuilt a property-data site 9 times overnight — jamesbrooksco · 2026-10-14
- Open-Generative-AI hits 30k stars: 600+ free models as a Sora/Veo alternative — matchaman11 · 2026-10-14
- Hunting old Linux gadgets to add to your mesh so agents can run stuff on them remotely — _Stocko_ · 2026-10-14