Getting started

Reviewing prompt coverage

Inspect tracked questions, topics, and collection settings, then propose questions to add or revise.

When to use this workflow

“Review our tracked prompts for gaps and duplicates.”

Before fetching

Establish audience, products, buyer stages, and reports in use. Read the complete prompt set before choosing high-impact corrections.

Quick reference

Use this sequence once you know the task. Read the numbered steps below for the reasoning and checks.

orc prompts list --json
orc prompts get YOUR_PROMPT_ID --json
orc answers list --prompt-id YOUR_PROMPT_ID --json

Prerequisites

Complete the CLI quickstart. These commands use CLI 0.5.0. Select the target workspace and use IDs returned by list commands.

Tracked prompts and their collection settings.

Read tracked questions and their settings. Look for topics with few answers or missing questions that prospective customers would ask.

1. Inspect tracked questions

Read question text and returned settings, then select the prompt IDs for this investigation. Fetch subsequent pages when needed.

orc prompts list --json

2. Read prompt settings

Replace YOUR_PROMPT_ID with the selected ID and inspect returned status, AI platforms, regions, and other settings.

orc prompts get YOUR_PROMPT_ID --json

3. Find relevant answers

Filter by the target brand or prompt and check answer timestamps. Choose an answer ID returned by this list.

orc answers list --prompt-id YOUR_PROMPT_ID --json

4. Verify and save the result

Produce proposed additions or revisions with reasons. Check the returned topic, status, AI platform, region, and other settings. Tracked prompts do not represent all market demand. If you change settings, record when they changed because they affect comparisons.

Record a reason for each prompt correction

Have the agent review each prompt for audience and buyer-stage coverage, branded-versus-general balance, duplicates, and ambiguity. Low answer volume alone does not make a prompt unnecessary. Record the prompt ID, quoted issue, impact, and suggested replacement for every finding.

Use taxonomy cleanup for classification problems and buyer-journey setup when creating a new prompt set.

When data is missing

If data is missing, check the workspace, collection period, filters, and pagination. Keep a failed request separate from an empty result. Use --help to inspect supported options.

If a failure remains, report it and verify the fix, including this workflow and the failed step.