Usage examples - ai-cmo.dev

Start with these copy-paste prompts, then follow the worked example to see the tool calls underneath.

Natural-language prompts

These prompts map directly onto MCP tools and are drawn from the install prompt text:

“Show my ai-cmo.dev score and scoreboard.”

Triggers aicmo_score + aicmo_scoreboard.

“List the latest recommendations and runs.”

Triggers aicmo_recs_list + aicmo_runs_list.

“Ask ai-cmo.dev what I should improve next.”

Triggers aicmo_ask with a prompt about improvement opportunities.

Worked example: accept a recommendation

This walkthrough shows what the assistant sees when a user asks to review and accept a recommendation.

Check the score

User: “Show my ai-cmo.dev score.”Assistant calls:

aicmo_score(workspace="my-brand")

Returns:

{
  "score": 68.2,
  "components": { "visibility": 55.0, "relative_sov": 72.0, "position": 68.0, "sentiment": 74.0, "volume": 65.0 },
  "date": "2025-06-15"
}

List recommendations

Assistant calls:

aicmo_recs_list(workspace="my-brand")

Returns:

[
  {
    "rec_id": "rec_abc123",
    "type": "gap",
    "topic": "brand positioning",
    "title": "Improve mention rate for brand positioning queries",
    "status": "proposed",
    "priority": 0.85
  }
]

Accept the recommendation

Assistant calls:

aicmo_rec_update(
  workspace="my-brand",
  id="rec_abc123",
  action="accept",
  expectedStatus="proposed",
  expectedVersion=1,
  idempotencyKey="unique-key-001",
  note="Starting work on brand positioning"
)

Returns:

{
  "rec_id": "rec_abc123",
  "from_status": "proposed",
  "to_status": "todo",
  "action": "accept",
  "version": 2
}

The recommendation is now in todo status and tracked in the workspace’s lifecycle.

​Natural-language prompts

​Worked example: accept a recommendation

Natural-language prompts

Worked example: accept a recommendation