Skip to scenario
AI-generated new docs

Make complete source evidence the gate for new documentation

A smaller docs team needs to document the bulk invoice editor for release 26.4. A thin ticket and a demo video leave room for plausible guesses. An agentic workflow needs source checks, content models, technical certification, and final documentation judgment.

See the comparison
The task in this example

Document the new bulk invoice editor for release 26.4.

Business goal
Increase efficiency
Business metric
Docs ready at release

A complete draft starts before generation

The same new feature needs documentation in both cases. The workflow changes whether incomplete sources reach generation and whether technical and editorial review have distinct gates.

A thin ticket becomes a whole documentation set

Generation begins without verified limits, roles, and error behavior.

The draft fills gaps with plausible steps that downstream assistants may repeat.

Inputs
Thin ticket and demo
Technical state
Uncertified claims
Placement
Proposed without corpus review
Review has to reconstruct the product evidence

The workflow waits for sufficient evidence

Sources pass the rubric before topics and their map are generated.

The product expert certifies accuracy; a documentation professional checks the collection and decides placement.

Inputs
Checked source package
Technical state
Expert certification recorded
Placement
Documentation professional approved
The release has a reviewed documentation set

Track the complete path to release-ready docs

This scenario covers a collection of new topics and its place in the existing corpus. It includes drafting, but the business result depends on certification, structure, placement, and publication.

  1. Prepare evidence

    Generation waits for the source package to pass.

    • Time to first draft
  2. Certify the collection

    Experts check technical completeness and accuracy.

    • Certified first time
  3. Finish for readers

    Documentation professionals review and place the topics.

    • Draft acceptance
  4. Release and observe

    Publish with the feature and inspect subsequent problems.

Business metric

Docs ready at release

Track whether the required collection is approved, placed, and available when the feature is released.

What to measure and how to interpret it

Start the clock at a defined intake milestone and record time waiting for missing sources separately. Compare features with similar documentation scope.

Time to first draft
Elapsed time from defined intake to a reviewable draft set; retain source-completion waiting time rather than excluding it.
Certified first time
Share of topics passing the product expert’s technical review on their first review cycle.
Draft acceptance
Share of generated content retained after substantive technical and documentation review.

Check the business result

Assess release readiness with total writer, expert, review, and rework effort. Check documentation-related tickets after release. A fast first draft is not the same as complete docs, lower cost, or a successful release.

Define what the workflow must know before it writes

Content models and source rubrics make expectations explicit. The source package must establish behavior; technical certification and final QA serve different purposes.

Release 26.4 / Bulk invoice editor
Generation intake rubric

Evidence required for a documentation set

Generation waits until source owners resolve missing behavior and scope.

Behavior
Supported actions and limits demonstrated
Context
Edition, role, release, and UI labels
Exceptions
Errors, partial failures, and recovery evidence
Content model
Required topic types and collection structure
Review gates
Product certification; terms/style; final docs QA; placement

Intelligent content

Maintain source rubrics, topic models, templates, and source rules. Generated topics must meet the same structure and reuse requirements as existing content.

Taxonomy

Use approved terms and curated classification values for edition, role, release, and feature. Rule-based checks enforce checkable terms and style rules.

A practical starting point

Choose one feature with cooperative source owners. Record the source gaps, technical corrections, final QA changes, and placement decisions across the whole workflow.

The model proposes content. Product experts certify behavior; documentation professionals decide usability, placement, and publication.

See how this scenario works
Lief
Lief
Amber
Amber
Dan
Dan
Explore the work with Intuitive Stack

What would make generated documentation ready for your release?

Intuitive Stack can help define source rubrics, content models, and review gates for a complete new-documentation workflow with your product and docs teams.

  • Generation-ready source evidence
  • Technical certification and final QA
  • Placement in the existing corpus