Skip to main content
Project To-dos gives each job one continuous list of required actions for everyone you have hired. Use it for onboarding steps, guideline updates you need acknowledged, training or quizzes, and work you want to review before an AI trainer continues. The list is living, not a one-time checklist. Anything you add appears on each assigned person’s list immediately, and you can include future hires automatically. Project To-dos is off by default for every job. Nothing is shown to you or to AI trainers until you enable it.

​
Enable Project To-dos for a job

  1. Open the job and go to the Settings tab.
  2. Find the Project To-dos section.
  3. Turn on Enable Project To-dos.
  4. Click Open Project To-dos to start building the list.
Enabling the setting adds a Project To-dos tab to the job for you and your team. AI trainers see their own to-do list in the manage-job view after a to-do is assigned to them. Turning the setting off hides the tab and all requirements again. Existing assignments and history are preserved, so re-enabling later does not lose anything.

​
Create a to-do

From the Project To-dos tab, click New to-do. Each to-do has:
  • Title and Instructions that explain what the person should do and why.
  • A completion method:
    • Worker marks it as done for simple confirmations.
    • Worker reads and acknowledges for guidelines and policy updates.
    • Worker completes a link (training, quiz, document) for external training or documents. Add the link in Link to complete.
    • An OpenTrain form or quiz built in the native builder. See Forms and quizzes.
    • Worker submits work for my review when you want to approve the result before it counts as done.
  • An optional category. Categories only group the list. They do not change who receives a to-do.
  • More options for urgency (Urgent, High, Normal), blocking behavior, and a due date in days.
Blocking options control whether the to-do gates project work:
  • Block project work until done
  • Block project work after the due date
  • Don’t block project work
Blocking to-dos only gate project work surfaces such as labeling. They never block an AI trainer’s time records, earned payments, messages, support, or disputes. When you edit a to-do later, use Publish an update. Updates create a new immutable version of the to-do. Existing completions and history are never rewritten, and you can choose whether people who already finished must complete the updated version.

​
Assign a to-do

When you save a to-do, you choose exactly who receives it:
  • Everyone — current people and future hires
  • Everyone currently on the project
  • Only future hires
  • Specific people…
Choosing an audience is always an explicit step — there is no default. Audiences that include future hires turn on automatic assignment: every person you hire on the job later receives the to-do without any extra step. The to-do card shows Includes future hires or Current assignments only so you can see the current behavior at a glance. Choosing Specific people… never includes future hires. From a to-do’s menu you can also:
  • Assign again / more people to reach additional people or require the to-do a second time.
  • Stop including future hires to turn off automatic assignment without touching existing assignments.
  • Remove this requirement to close one person’s open obligation. A reason is required, and the person’s history is kept.
  • Archive… to retire the to-do. Open copies close, and history is kept.

​
Review submitted work

To-dos that require your review appear in the Review tab of the Project To-dos panel once an AI trainer submits. For each submission you can:
  • Approve to mark it complete.
  • Return to worker with feedback about what needs to be corrected. The AI trainer sees your feedback and can resubmit.
  • Waive to drop the requirement for that person. A reason is required.

​
Track completion

Use People & results on any to-do to see per-person status: who has not started, who is in progress, who submitted, and who is done, waived, or closed. You can filter by person. Use Export CSV from the Project To-dos panel to download the full list with per-person statuses for reporting.

​
What AI trainers see

Each AI trainer sees only the to-dos assigned to their own contract, in a Project To-dos tab of their manage-job view. Their list separates to-dos that are required before project work, other open to-dos, and items waiting on your review. See Complete project to-dos for the AI trainer side.

​
Automate with the API, CLI, or MCP

Agents and scripts can manage the same list through the developer surface, with mandatory audience selection, no-write previews, and signed confirmation for broad assignments. Start with the Project To-dos developer concept.