Tasks are how you produce a Key Deliverable. Each task generates a specific part of the deliverable, so you build the document by selecting and running the tasks you need rather than generating everything at once.
Selecting tasks
Within a Key Deliverable, the work is organised into groups, each holding its tasks. You browse these and select the tasks you want to produce. You only need to select the tasks relevant to your project. Your selection is the request, and Cognibuild produces what you choose and nothing beyond it. You can add more tasks at any time as your project develops.
Running tasks
Once you have selected your tasks and assigned your supporting documents, you run them. When you run a task, Cognibuild reads your assigned documents and generates structured content for that part of the deliverable. The generated output appears in the deliverable, where you can review it.
Rerunning tasks
After reviewing a task's output, you may want to refine it. Rerunning a task generates an updated version of that section. This is useful when you have added or updated supporting documents, refined your project context, or simply want to improve the result. Each rerun produces a fresh output for that section, without affecting the rest of the deliverable.
Building up the deliverable
Most deliverables are produced through several rounds of this: run a task, review the output, adjust your documents or selection if needed, and rerun until the section reflects what you need. Because tasks generate individual sections, you can build and refine a deliverable gradually, section by section, as your project information becomes clearer.
Running and rerunning tasks uses credits, since each one involves AI processing. Selecting tasks, reviewing output, and editing manually do not.