Teams Quickstart
Create a board, hire an AI employee, assign a clear task, and review the first result in PromptRails Teams.
Best for
People assigning work to AI employees and reviewing their results
Start with one small, reviewable task. This walkthrough uses a weekly research brief: an employee gathers context and produces a draft for a person to review.
1. Open your workspace
Sign in to Teams with your PromptRails account, or switch to Teams from Console. Select the workspace where the work belongs. You need a usable model and enough workspace balance to run an AI employee; an admin can check usage and billing if a run cannot start.
2. Create the board
Open Boards. Create a board yourself or choose a template whose stages match your process. To have Teams help with the setup, choose New → Workflow, described in words and describe the outcome, the people who review it, and the tools it needs. Review the proposed setup before using it.
For a research brief, use stages such as Inbox, Research, Review, and Done. In the board settings, write the shared instructions: cite sources, distinguish evidence from assumptions, and leave a draft for review.
3. Choose an employee
Open Team, or choose New → Hire an AI employee. Review a specialist’s role, included skills, and playbook before hiring. Open the hired employee’s profile to see what it can do. An admin can connect the tools and knowledge it needs.
Hiring an employee does not connect your external accounts automatically. If the task needs an integration, have an admin connect it and grant the employee access before assigning the work.
4. Add a clear task
Create a card with a title and a brief. For example:
Draft a weekly competitor research brief for our product team. Use the sources attached to this card. Include three changes, links to the evidence, and open questions. Label assumptions. Leave the brief for human review.
Attach relevant files, set a due date if needed, and assign the employee. Assignment can start an AI run; use a task you are ready for it to work on.
5. Follow progress and review
Open the card to read the discussion, run progress, and any deliverables. If the employee needs a decision, open Needs you and inspect what is being requested before approving or denying it.
Read the result and its sources. Accept useful work or request specific changes. A completed model run alone is not a quality verdict: the output still needs to meet the brief.
6. Make the process repeatable
Once the first result is useful, refine the board instructions and employee guidance. Then consider a recurring workflow or automation. Keep approval gates where the employee needs a person’s decision.
Continue with Boards and cards, AI employees, and Approvals and automation. For runtime debugging, open the same workspace in Console and inspect its traces.