Use cases › IT / internal ops

ShipStation "Deckhand": how the agent works and how to build one like it

Summary: ShipStation built "Deckhand," a Slack IT help-desk agent that opens Jira tickets and answers 94% of after-hours questions, per the source.

Source: https://aws.amazon.com/blogs/machine-learning/how-shipstation-global-built-an-ai-it-help-desk-on-amazon-bedrock-agentcore/ source read via search extract

Why this use case works

IT help desk is one of the easiest first agent projects: questions repeat, answers live in existing docs, users are employees (more forgiving than customers), and the fallback is just a ticket. After-hours coverage is where the value is most visible. (Our analysis.)

Agent pattern: Internal ops agent

An internal-ops agent living in a chat tool employees already use (Slack, Teams, WhatsApp), with read access to documentation and write access limited to creating tickets or records. Start read-only and widen permissions as accuracy is proven.

How a smaller team could replicate it

Any company on Slack or Teams can index its IT and HR docs, let the agent answer and create tickets in Jira, Linear or a help-desk tool, and review unanswered questions weekly to fill doc gaps. (Our analysis, not a claim made by the source.)

Want a rough business case first? Try the AI agent ROI calculator, or ask us to scope it.

Risks and guardrails

Risks: confidently wrong answers from outdated docs and over-broad permissions. Guardrails: cite the doc behind each answer, keep write access limited to ticket creation, and review unanswered or downvoted questions weekly. (Our analysis.)

What to measure

After-hours answer rate, tickets auto-created versus manually filed, time to resolution, and employee thumbs-up rate.

Related IT / internal ops deployments

Part of the AgentStack Atlas catalog of 52 sourced agentic AI deployments. Spotted an error? Email agentstackatlas@mail.grokbot.com.