OpenAI Engineer Demonstrates Dots for Engineering On-Call Workflows

An OpenAI engineer describes how Dots can operate as an always-on cloud agent with its own cloud computer, access to a local computer, and an independent identity across services such as Slack. The demonstration centers on configuring Dots for engineering on-call workflows, including alert monitoring, classification, and investigation coordination.

■ Dots setup and Slack integration
- Connecting a Dot to Slack to give it an identity
- Using Composer to set an automation that monitors a channel, classifies incoming alerts, and starts investigation threads

■ On-call incident example
- Receiving an alert on a phone, acknowledging it, and checking the Dot's progress in the ChatGPT app
- Dot completing an investigation, identifying the issue, and creating a Codex app for continuing the thread
- Sending messages to a laptop and allowing work to continue from either phone or laptop

■ Additional workflow examples
- Triaging incoming notes in a specified channel
- Mentioning the Dot in Slack to create pull requests and work on them in Slack

This is suited to engineers and developers exploring AI agents for incident response, alert triage, and Slack-based engineering workflows. Viewers can see an example of configuring an always-on agent and understand the types of tasks it can coordinate; the transcript does not cover implementation details or pricing.