An AI Agent That Releases Blocked Query Workgroups – and Teaches Users Not to Repeat the Mistake

An AI Agent That Releases Blocked Query Workgroups – and Teaches Users Not to Repeat the Mistake

#ai agents #athena #automation #aws #jira
team develeap
September 17, 2026

Built with Mobileye’s cloud support team

The Problem

Data and algorithm engineers run heavy SQL queries over massive datasets in AWS Athena, organized into workgroups per team. When a workgroup exceeds its resource quota, AWS disables it – and every user on that team is blocked from running queries until someone manually re-enables it.

Every block becomes an urgent support ticket. The fix is technically simple, but it needs immediate human attention, often at inconvenient hours.

The Solution

The workgroup agent detects these tickets the moment they open and runs the release end to end:

  1. Recognizes an enable/disable request from the ticket language.
  2. Extracts the exact workgroup name and required action.
  3. Executes the AWS operation to re-enable the workgroup.
  4. Replies to the user in Hebrew or English confirming completion – and adds an educational reminder to check for inefficient queries, with a link to the support team for optimization help.
  5. Asks one focused follow-up question if the workgroup name is missing, instead of a round of back-and-forth.
  6. Labels, assigns, and updates the ticket status.

The Results

Blocks that waited for a human are resolved within minutes, regardless of who’s available. Every release comes with proactive guidance, leading to a measurable drop in repeat blocks on the same workgroups. And handling is consistent no matter which engineer would have picked up the ticket.

Why It’s Worth Talking About

A classic operations quick win: a simple, focused action automated to save response time for blocked users – with the agent delivering extra value through personalized guidance.