> For the complete documentation index, see [llms.txt](https://docs.ratl.ai/guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ratl.ai/guide/features/agentic-interface.md).

# Talking to Agents: A Comprehensive Guide for Interactions

<figure><img src="https://3939944954-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FefV6n9Vxvs062CsQhtKk%2Fuploads%2FQIfOcCCrKywy71Iy0rot%2Fimage-1.png?alt=media&amp;token=b80502c8-5aa5-4ad8-89b6-61c3f51ad09c" alt=""><figcaption></figcaption></figure>

### Key Benefits

* **Automated Testing Assistance**: Agents can handle complex testing tasks like functional tests, load tests, end-to-end integration tests and accessibility testing.
* **Seamless Communication**: Interact with sub-agents like **INGA**, **CHASE**, and **CASEY** to streamline testing and reporting.
* **Multi-tasking Support**: Perform various tasks such as project creation, script generation, execution, and report analysis all via chat.
* **Improved Efficiency**: Agents can automate repetitive testing tasks and provide detailed test reports, helping you save time and effort.

***

### Preconditions

Before you interact with agents, ensure the following:

* You have a valid **ratl.ai** account and necessary permissions to run tests.
* Workspaces are created or ready for testing. If not, you can create them via the agent.
* API specifications, Postman collections, or cURL commands are accessible for API testing.
* Understand the types of testing you want to perform, e.g., **functional**, **load**, **end-to-end integration or** **accessibility testing**.

***

### How to Interact with Agents

**1. Initiating a Conversation**:

* Start by greeting or asking RATL for assistance.

  **Example**:

  ```
  Human: Hi RATL, what are your capabilities?
  ```

  <figure><img src="https://3939944954-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FefV6n9Vxvs062CsQhtKk%2Fuploads%2FKoGDpwfc1KYGFyioz3rd%2Fimage-2.png?alt=media&amp;token=8db3f035-9aa3-4c48-83bd-372e5b3ad389" alt=""><figcaption></figcaption></figure>

**2. Create a Workspace**:

* You can create or select an existing workspace to get started.

  **Example**:

  ```
  Human: Create a workspace called "Automation-Station".
  ```

  <figure><img src="https://3939944954-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FefV6n9Vxvs062CsQhtKk%2Fuploads%2FNBwf1y9JMZydqJ685TxL%2Fimage-3.png?alt=media&amp;token=b7728596-f0bc-4716-a4ef-0ea28e0bc4f2" alt=""><figcaption></figcaption></figure>

**3. Start Testing**:

* Use Postman collections, cURL commands, or other inputs to run tests.

  **Example**:

  ```
  Human: Proceed with functional testing using this curl: curl -X 'GET' 'https://reqres.in/api/users/' -H 'accept: application/json'.
  ```

**4. Perform Load Testing**:

* Provide the API endpoint to run load tests.

  **Example**:

  ```
  Human: Let's perform a load test using the curl: curl -X 'GET' 'https://reqres.in/api/users/' -H 'accept: application/json'.
  ```

**5. Request Test Reports**:

* After executing tests, you can request reports.

  **Example**:

  ```
  Human: Get the report for the load test.
  ```

**6. Generate and Run Integration Flows**:

* Provide Postman collections to generate integration flows.

  **Example**:

  ```
  Human: Perform integration tests using this Postman collection: {collection data}.
  ```

**7. Perform Accessibility Testing**:

* Ask the agent to run accessibility tests for websites.

  **Example**:

  ```
  Human: Perform accessibility test on https://www.jiomart.com/.
  ```

**8. Integrate ADO(Azure DevOps) to a Workspace**:

* Provide the agent with the ADO Personal Access Token (PAT), ADO organization name, associated email address, and workspace name.

  **Example**:

  ```
  Human: Integrate ADO to the Automation-Station workspace with the following credentials: {ADO PAT}, {ADO organization}, {ADO email}.
  ```
* The agent will display a list of projects associated with the provided ADO credentials, simply select one to complete the integration.

  **Example**:

  ```
  Human: Integrate this ADO project: {ADO project name}.
  ```

**9. Integrate Jira to a Workspace**:

* Provide the agent with the Jira API Token, Jira organization name (subdomain name), associated email address, and workspace name.

  **Example**:

  ```
  Human: Integrate Jira to the Automation-Station workspace with the following credentials: {Jira API Token}, {Jira organization}, {Jira email}.
  ```
* The agent will display a list of projects associated with the provided Jira credentials, simply select one to complete the integration.

  **Example**:

  ```
  Human: Integrate this jira project: {jira project name}.
  ```

**10. Integrate Slack to a Workspace**:

* Provide the agent with the Slack Bot Token and workspace name.

  **Example**:

  ```
  Human: Integrate Jira to the Automation-Station workspace with the following credentials: {Slack Bot Token}.
  ```
* The agent will display a list of slack channels associated with the provided slack token, simply select one to complete the integration.

  **Example**:

  ```
  Human: Integrate this slack channel: {slack channel name}.
  ```

**11. Generate API Keys for CI/CD Integration**:

* Ask the agent to create a new api key and provide it with an expiry date.

  **Example**:

  ```
  Human: Create a new api key that expires after 30 days.
  ```

***

### What to Do

* **Be clear with instructions**: Provide clear and concise commands like "perform a load test" or "create a workspace."
* **Provide correct inputs**: Ensure your API specifications, cURL commands, or Postman collections are correctly formatted.
* **Use prompts**: When modifying scripts or configurations, use prompts or the built-in code editor to make adjustments.
* **Review agent responses**: After performing tasks, review logs, reports, and execution statuses provided by the agents.
* **Monitor test status**: Keep track of the test statuses and logs through links provided by the agent.
* **Ask for help**: Don’t hesitate to ask the agent for further actions like generating reports or fetching statistics.

***

### What Not to Do

* **Avoid incomplete inputs**: Make sure to provide complete cURL commands, Postman collections, or other specifications. Incomplete inputs may lead to test failures.
* **Don't skip workspace creation/selection**: Always ensure a workspace is created or selected before starting testing.
* **Don’t run multiple tests without reviewing**: Ensure you review the results of a test before starting another one to avoid confusion.
* **Avoid overly complex commands in one interaction**: Break down your tasks into smaller, manageable steps for smoother interaction with the agent.

***

### Postconditions

After interacting with the agent, expect the following outcomes:

* **Workspace Creation**: The workspace is created and ready for testing.
* **Test Execution**: Test cases are executed, and logs are provided for review.
* **Reports**: Detailed reports are generated and made available for download.
* **Error Handling**: Any issues with scripts or testing are highlighted, and the user is informed to correct them.
* **Task Completion**: Tasks are completed, and the agent confirms the result.

By interacting with the **RATL Agent**, you can streamline your testing and automation workflows efficiently, all while having clear visibility into the test execution and results.