Canonical page: https://www.kuudo.com/features/skills/amazon-ads/marketing-cloud/

Amazon Ads Skills v0.5.2 Reference

# Turn AMC Questions Into Answers and Audiences

Ask in plain English — new-to-brand, ad frequency, customer lifetime value, off-Amazon conversions — and the skill picks from 50+ proven AMC playbooks, writes validated SQL, and returns privacy-safe answers or DSP-ready audiences. The Amazon Marketing Cloud Skill reads the matching playbook before writing SQL, and it accepts AMC terms only when you authorize it.

Grounded before execution

[Try the starter prompt](#starter-prompt)[Connect Skills](/docs/quick-start/chatgpt-skills/)

Example output visualization: AMC workflow

AMC workflow

skill output

1

Plain-English question

2

Atlas playbook

3

AMC SQL

4

Execute workflow

Answer or DSP audience

Primary input

A plain-English question

Primary output

Validated AMC SQL

[

Powered by

Amazon Ads MCP

](/features/amazon-ads-mcp/)[

Source contract

amazon-ads-marketing-cloud

PolyForm-Shield-1.0.0

](https://github.com/KuudoAI/skills/blob/674e68adc1c0242e5ea440583f82d26c93c86b1d/skills/amazon-ads-marketing-cloud/SKILL.md)

Is this the right Skill?

## Use it for a clear operating job.

Best for

*   New-to-brand, reach, and frequency questions
*   Custom attribution and cross-channel overlap
*   Rule-based or lookalike audiences for DSP

Use another path when

Questions that can be answered directly from a standard Amazon Ads report without AMC SQL.

Starter prompt

## Start with the task, then add your scope.

Copy this into a connected client and replace the account, date, ASIN, or threshold details with yours.

```text
Use AMC to find how many ad exposures it takes before our DSP and Sponsored Products ads stop adding purchases, then build a frequency-capped audience I can activate in DSP. Ground it in the playbook, show the SQL and privacy constraints, and run after I confirm the scope.
```

The Skill will request any required context or approval before guarded actions.

The outcome

## A defined result, not a generic answer.

A question that used to need a data team comes back as an executed AMC workflow: validated SQL, privacy-safe results, and — when the job is an audience — a segment ready to activate in DSP.

Inputs

*   A plain-English question
*   AMC instance and advertiser
*   Date range, ASINs, and campaign scope

Outputs

*   Validated AMC SQL
*   Workflow execution
*   Analysis or DSP-ready audience

Workflow

## The playbook your agent follows.

1.  01
    
    ### Retrieve the playbook
    
    Match the question to one of 50+ AMC playbooks — attribution, frequency, lookalikes, lifetime value — before any SQL is written.
    
2.  02
    
    ### Resolve the instance
    
    Set advertiser context and choose the correct AMC instance.
    
3.  03
    
    ### Create and execute
    
    Create the workflow, validate the query, and run it with explicit parameters.
    
4.  04
    
    ### Interpret the result
    
    Download results, check privacy thresholds, and deliver the analysis or DSP-ready audience.
    

Guardrails

## What keeps the workflow dependable.

Playbook retrieval precedes SQL generation

Privacy and audience-size constraints remain explicit

Results preserve workflow and instance provenance

Connected systems

## The context and tools it coordinates.

Amazon Ads MCP

Amazon Marketing Cloud

Agent Atlas

Keep building

## Related workflows and field guides.

[All Amazon Ads Skills →](/features/skills/amazon-ads/)

[Related SkillAmazon Ads reporting](/features/skills/amazon-ads/reporting/)[Related SkillSponsored Products search terms](/features/skills/amazon-ads/sponsored-products-search-terms/)[Field guideTurn Amazon Marketing Cloud Workflows into Skills](/guides/amc-agent-workflows/)

amazon-ads-marketing-cloud

## Give your agent the complete playbook.

Connect Skills once, then start with the prompt above whenever this job comes up.

[View prompt](#starter-prompt)[Connect Skills](/docs/quick-start/chatgpt-skills/)