Getting started

​
Set up prompts across the buyer journey

Create a balanced set of awareness, comparison, decision, and brand-evaluation prompts.

​
When to use this workflow

“Start monitoring this brand from early research through purchase decisions.”

​
Before fetching

Confirm the brand, audience, products, region, language, platforms, and schedule. Read the service model and region catalogs and the existing prompts before the agent drafts candidates.

Complete the CLI quickstart and confirm the workspace with orc workspace current --json. Replace example IDs, URLs, dates, and settings. Apply only approved mutations.

​
1. Check the channel and region catalogs

Get the current channel and region catalogs from the service. Use the stable channel id and region code returned by those commands. Catalog membership does not guarantee access under your plan or credential.

orc channels list --json
orc regions list --json

Run the measurement flow in this order: channels list → prompts create --platforms <channel> → runs create --model-channel-id <channel> --wait → answers list.

​
2. Read the existing setup

Read brands, topics, and prompts, then build a stage-by-intent matrix. Ground product-specific questions in product information.

orc brands get YOUR_BRAND_ID --json
orc topics list --brand-id YOUR_BRAND_ID --json
orc prompts list --json

​
3. Draft candidates and validate each prompt

Have the agent propose questions for missing stages and select the approved candidates. Preview resolves configuration without saving. Replace YOUR_REQUEST_KEY with a unique key for this request.

orc prompts create --topic-id YOUR_TOPIC_ID --text 'YOUR_APPROVED_QUESTION' --platforms YOUR_CHANNEL_ID --region-id YOUR_REGION_CODE --preview true --idempotency-key YOUR_REQUEST_KEY --json

​
4. Save approved prompts and read them back

After checking the resolved settings, create with preview omitted and a new request key. Read the returned ID; use tags update when assigning existing tags. Include retained tags because the operation replaces the set.

orc prompts create --topic-id YOUR_TOPIC_ID --text 'YOUR_APPROVED_QUESTION' --platforms YOUR_CHANNEL_ID --region-id YOUR_REGION_CODE --idempotency-key YOUR_CREATE_REQUEST_KEY --json
orc prompts get YOUR_CREATED_PROMPT_ID --json
orc prompts tags update YOUR_CREATED_PROMPT_ID --tag-ids YOUR_TAG_IDS --config-revision YOUR_CONFIG_REVISION --json

Run the saved prompt on the channel, then read the answers back with the same channel filter.

orc runs create --prompt-id YOUR_CREATED_PROMPT_ID --model-channel-id YOUR_CHANNEL_ID --region-id YOUR_REGION_CODE --idempotency-key YOUR_RUN_REQUEST_KEY --wait --json
orc answers list --prompt-id YOUR_CREATED_PROMPT_ID --platform YOUR_CHANNEL_ID --json

​
Deliverable

Deliver stages, intents, question text, settings, and created IDs, plus remaining coverage gaps and excluded candidates.

On fetch failure, retain the error and request scope. Do not count empty data, incomplete pagination, or missing permissions as a completed investigation.

​
Next steps

Workflow index · Reviewing prompt coverage · Clean up topics and tags · Set up shopping prompts from products and personas

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