AI Agent Use Cases
Views —

AI Agent Use Cases

Added today

44All 189

Prompt to Prepare a Shopping Cart from a Recipe Photo

The author reports that Grok Bot receives a recipe photo, organizes the ingredients and quantities, and creates a shopping list by excluding ingredients already available at home.

Grok Bot
Life & Hobbies

1–20 of 189

Dev

Agent to Investigate Operations Alerts Logs and Code and Create a Fix PR

Carvana has shared a workflow where the author investigates logs and code related to operations alerts, creates a fix proposal, and then has it reviewed and verified by a human.

Claude

How it works

  1. The author connects the alert context, logs, and repository access, and assigns reviewers and deployment managers.
  2. Claude investigates the logs and code based on the alert context, creates a PR, and after human review and deployment, the normal state is verified.

Results

Carvana reports that they operate a workflow where Claude Tag investigates operations alerts, creates a fix PR, and after human review and deployment, the logs are checked.

Notes

This is an explanation of how Carvana uses Claude Tag in operations. The workflow involves a human reviewing and deploying the fix PR, then verifying the system state.

Source

Carvana / Claude 고객 사례 Original post ↗

Example prompt · Claude

Goal: Investigate logs and code from operations alerts and create a fix PR.

Inputs/Environment: [Sample operations alert], [readable logs], [test repository], [review procedure].

Reference original: https://claude.com/customers/carvana

Based on the published workflow, prepare an implementation or task procedure that can be followed in a test environment. Do not guess at internal code or settings not disclosed in the original; if alternative connections are needed, please indicate so.

Connections and materials to prepare: Connect the alert context, logs, and repository access, and assign reviewers and deployment managers.

Agent actions to implement: Claude investigates the logs and code based on the alert context, creates a PR, and after human review and deployment, verifies the normal state.

Scope of actions: This is an explanation of how Carvana uses Claude Tag in operations. The workflow involves a human reviewing and deploying the fix PR, then verifying the system state.

The deliverable is an investigation of the operational issue and a fix PR to be reviewed. Please also organize the required setup steps and actual execution methods. After checking the results returned by the tool, select the next task, and separately note any materials that could not be accessed or tasks that could not be executed.

Verification: Check whether the root cause hypothesis matches the logs, whether the fix prevents the issue from recurring, and ensure that no production deployment occurs before review. Clearly distinguish the inputs used for verification, the actual outputs, and any items not yet confirmed.