Before
An analyst submitted a search query, waited for the search job to complete, and checked its status manually. They then retrieved the results and copied the relevant evidence into the requesting workflow or record.
How it works
- Accept the submitted search query and create a search job.
- Wait briefly, then check the job status before retrieving its results.
- Count and package the returned evidence rows with the final status for the parent workflow.
Workflow details
- Search submission: The workflow receives a search query from a parent process and creates a corresponding search job; in the configured website demonstration, the job-creation response is replayed locally.
- Job monitoring: After a short wait, it checks the search-job status so the workflow can proceed once the search is ready.
- Evidence retrieval: It requests the available result rows for the completed job, including directory or other evidence records returned by the search.
- Response preparation: The workflow normalizes the status, counts the returned rows, and packages the results into a structured response for the calling workflow.
- Operational handoff: The parent workflow receives a consistent search outcome without an analyst repeatedly checking job progress or copying individual result sets; the demonstration uses synthetic responses to replay the external reads and writes.
What changed
Search status checks and result retrieval now happen in a reusable workflow step, with the final status, count, and evidence rows returned directly to the parent workflow. Human involvement shifts from managing the handoff to using the returned evidence for the next decision.
Work removed
Manually polling search jobs, checking whether a search was ready, retrieving result sets, counting returned rows, and copying evidence into the requesting workflow.
Expected outcome
A consistent evidence handoff for each submitted search, automating two recurring manual action categories: job polling and result copying.
