Agent builder
From a request in Portuguese to the agent: interpret, review the powers one by one, rehearse without executing and create, always starting at suggest.
What you find on this screen
A portrait of the screen, read from the screen itself — it changes when the screen changes.
Tabs
What it may read and doWhat would you do today?Versions of this agent
Available actions
Read my requestReinterpret this textSave the draftRun nowOpen new versionWhat would you do today?Set its autonomy levelSee its decision trailOpen the trail (full list)see the list behind this number
The list shows
AutorizoWhatTipoScreen that authorizesAuthority levelRequest in the textWhy it was left outEtapaWhat would happenVersionStatusPedido+2
Start here
From text to agent, no skipped steps
Write the request and click Interpret: Tenkai proposes name, goal, trigger, policy and caps. Review the powers on their tab (each shows the screen that authorizes it), run the rehearsal without executing and save the draft. Only "Create the agent" creates it — and it starts in suggest mode.
How to
Save, browse and undo
Save stores without leaving the form (Ctrl+S); Save and close returns to the list; the header arrows move from one record to the next. Delete sends the record to the list’s trash, from where Undo brings it back.
Standard list and form
The screen became the standard list: search, filters, saved views, columns, export and trash with Undo. Click the record (or New) to open the form, with the detail tabs when there are any.
Start from a template or from Tenkai's bubble
The four ready templates (Financial Analyst, Production and Purchasing Planner, Anomaly Detector, Sales Assistant) fill in the request and interpret it right away; you check the powers and publish. The "Repeat every Monday" button on any Tenkai answer lands here with the request built from the question. Publishing is still explicit.
Rules
Run now, without waiting for the schedule
After creating the agent, the Run now button makes it work right away and records the result in the Decision trail, just like a scheduled run. Use it to see the agent working before its first scheduled time. It does not run if agents are turned off on the platform or if this agent is paused in Autonomy; in those cases the screen says which was the reason.
Was this article helpful?
In the same module: Digital employees23
Sprint developerAnchor screen of the detail dictionary (allocated devs). Does not appear in the menu.Workflow transitionAnchor screen of the detail dictionary. Does not appear in the menu.Caixa de entrada dos funcionários digitaisO que os agentes estão esperando de você, reunido num lugar só e separado por urgência.Accountant's cockpitThe accountant's dashboard: what's pending, what's due, and what needs approval, per company in the portfolio…My hiresThe marketplace transactions with the flow open: the ones I hired (I am the buyer) and the ones that hired me…Proprietary credit scoreThe credit risk of each drawee (customer) measured by the OBSERVED BEHAVIOR in your own operation — not a bur…Auditor package (SOC 2)The package handed to an auditor: the Segregation of Duties (SoD) matrix with the violations found — profiles…Partner's userAnchor screen of the detail dictionary. Does not appear in the menu.
Tenka Academy
Explore the AcademyStep-by-step learning paths for your industry, inside the platform.