Outputs

A field can produce multiple outputs — the primary value plus AI-driven derivations, each emitted to its own case-data key. Outputs surface in the Case Data panel under the Computed bucket. They do NOT appear as separate dashboard columns — the column resolver expands template fields but not their secondary outputs.

Last updated

Every field has at least one primary output (the case-data value it produces) and can declare additional outputs for derived values — AI prompts, computed expressions, or context bindings.

What you need

  • A workspace with at least one template
  • Owner or member role to edit templates

What an output is

Open any field's editor on a template. Each field exposes an Outputs section. Outputs can be:

  • Input — the user's form input (for question fields).
  • Static — a fixed value.
  • Zone — extracted from a read zone you draw on the source document.
  • Prompt — derived from a natural-language AI prompt running over the rest of the case data.
  • Computed — derived from an AST expression referencing upstream output keys.
  • Binding — read from email metadata, the email body, system context (today's date), or the company profile.

Each output gets its own case-data key, its own value type, and is rendered independently in the Case Data panel.

Where outputs surface

Outputs appear in the Case Data panel of the case detail. Primary outputs render under their field's bucket (User Input, Extracted, or Static); derived outputs (prompt/computed/binding) appear under the Computed bucket.

Outputs other than the primary are not dashboard columns. The dashboard table column resolver (resolveWorkspaceTableColumns) walks the workspace's templates and expands their top-level fields. Secondary outputs do not appear as separate columns. If you want a derived value as a sortable dashboard column, model it as its own top-level template field instead.

When to use a derived output vs. a separate field

  • Use a derived output when the value is derived from other case data — totals, deltas, conditional categorizations, formatted versions of an existing field.
  • Use a separate extracted field when the value lives in the document and the AI just needs to find it. Don't ask a derived output to extract; that's what zone or prompt outputs on the source field are for.

Step by step

  1. Open the template in the studio.
  2. Click the field whose value you want to derive.
  3. In the Outputs section, add a new output.
  4. Pick the kind — the picker for derived outputs offers Computed, AI instruction, Static value, and Binding — and fill in the rule (prompt text, expression, static value, or binding source + path). The primary output's kind is set implicitly from the field's source + whether a read zone is attached; only derived outputs expose this picker.
  5. Click Save.

The next autofill or re-extraction populates the output's value on every relevant case.

Tips

We use cookies to keep you signed in and improve the product. See our Cookie Policy.
Manage preferences