Skip to main content

Automatically Registering Review Findings to Lightning Review

Overview​

This is a sample of performing reviews on Word design documents using a workflow, aligned with review criteria.

The design document to be reviewed and the review criteria are read, and the agent is repeatedly executed in units of review criteria × chapter.
The extracted review findings are output to Lightning Review.

In this sample, the following are executed sequentially as a workflow:

  1. Obtain the document to be reviewed
  2. Obtain the review criteria
  3. Repeatedly execute the review
  4. Output to Lightning Review

This demonstrates how to proceed with reviews of Word documents according to review criteria, combining tool integration and an AI agent.

What this sample can do​

  • Integrate with Lightning Review and register extracted findings in Lightning Review
  • Load a review perspective list file and conduct reviews using the review perspectives
  • Break down a Word-format design document into chapters and conduct reviews
  • Comprehensively and repeatedly perform reviews using combinations of chapters and review perspectives

Target Users​

This sample is primarily intended for the following users:

  • Those considering integrating Lightning Review with AI
  • Those who want to streamline and standardize design document reviews using AI
  • Those who want to experience iterative execution using workflows

Contents Included​

This sample includes the following data as an example:

TypeContent
Workflow DefinitionReview Word design document by chapter and perspective and output findings to LR.yaml
Agent DefinitionWord design document review_LR findings JSON generation.yaml
Input: Review perspective filereviewPointList.json
Input: Word design documentAPP layer component design document.docx
DocumentationThis README

Using Sample Data​

Prerequisites​

  • The data structure of the review perspectives is the same as the input in this sample.
  • The Lightning Review file for outputting review findings is open.

Input/Output​

  • Input: Review perspective list file, design document in Word format
  • Output: List of all review findings file (Lightning Review format)

Procedure​

  1. Open the Lightning Review file for outputting review results.
  2. On the Agentiqs workflow screen, click "Review Word design document by chapter and perspective and output findings to LR".
  3. On the screen for "Review Word design document chapter by chapter and perspective and output findings to LR," click "Execute."
  4. A dialog box will appear. Enter the "Review Perspective List File Path" and "Word Design Document File Path," then click "Execute."
  5. Findings will be output to the Lightning Review file.

Target Operation​

  • Workflow: Review Word design document chapter by chapter and perspective and output findings to LR
  • Agent: Word design document review_LR finding JSON generation

Constraints​

  • This sample reviews by inputting data into the agent on a chapter-by-chapter basis. Therefore, accuracy may decrease if a chapter is very large. In that case, customize the scope of input to the agent, such as by section.
  • Review perspectives must be available in the format expected by the workflow and agent. If the structure or content differs, the workflow will need to be modified.

Reference: Workflow Execution Flow​

This sample demonstrates how to break down the process step by step using a workflow, and how to combine the tool and AI agent to execute a series of review processes.

  1. Read the review perspective list file and format it into a loopable structure for each perspective.
  2. Read the design document in Word format and format it into a loopable structure for each chapter.
  3. Input the information sequentially into the review agent for each chapter and each review perspective, and extract the review findings.
  4. Register the review findings in Lightning Review for each chapter and each review perspective.
Last updated on