Set the outcome
Describe what done looks like—one task, a recurring workflow or an ongoing responsibility.
How it works
From intent to finished work.
Start with a clear outcome. Your Dot organizes the work, uses connected tools and keeps you in control of key decisions.
Connect its workspace
Give the Dot access to the relevant apps and context through the available plugin ecosystem.
Let it operate
The agent reasons, browses and carries out steps from its dedicated cloud computer.
Review and direct
Follow progress from ChatGPT, Slack or Microsoft Teams and step in when judgment is needed.
Control the operating boundary
Autonomy works best with clear limits.
Before assigning ongoing work, define the sources a Dot may use, the actions it may take and the decisions that must return to a person.
Set permission scope
Connect only the tools and information the workflow genuinely needs.
Define check-in points
Choose when progress should be summarized and when work should pause for review.
Measure the outcome
Use completion criteria that are observable, specific and easy to verify.
Keep judgment human
Reserve sensitive, high-impact or ambiguous decisions for an accountable person.
Start with one outcome, not every process.
Pick a workflow with clear inputs, a defined finish line and an obvious moment for human review.