Manage scenarios

The latest service changes have not yet been reflected in this content. We will update the content as soon as possible. Please refer to the Korean version for information on the latest updates.

Available in Classic and VPC

A scenario is the process through which the model thinks and examines to provide appropriate answers to the user's questions. Through collecting scenarios and reviewing data, you can edit your model's thought process.

Scenario screen

Collect scenarios

To collect scenarios, follow these steps:

1. On the NAVER Cloud Platform console, click Services > AI Services > CLOVA Studio in order.
2. Click the My Product menu, and then click the [Go to CLOVA Studio] button.
3. On CLOVA Studio, click the Skill trainer menu.
4. Click the [Collect scenario] button at the top of the screen.
5. Select the skill set and engine from the left of the screen, and enter the user query.
   • Skill set: select a skill set in which to collect scenarios
   • Included skills: a list of skills included in the selected skill set. You can only use skills whose task status is "completed"
   • Engine and version: select the language model to use. The default language model is selected by default
   • Call option (optional): enter a fixed value when calling a user query. For more information on writing a call option, see Write call option.
   • User query: the field for entering user queries. You need to select a skill set to activate this field
6. Click the [Run] button.

Edit scenario data

To review and edit the collected scenario data, follow these steps:

1. Run Collect scenario.
2. Review the model's thoughts in the planning data area.
3. To edit one of the steps in the planning data area, click each relevant field.
   • For more information on how to edit content in the planning data area, see Planning data.
4. Once the step is edited, click the [Apply] button.
   • The step is regenerated.
5. To edit data in the skill data area, click each relevant field.
   • For more information on how to edit content in the skill data area, see Single skill data.
6. When you are finished with editing, click the [Apply] button.
   • The skill data will be rerun.
7. When the final response field is activated in the planning data area, click the [Complete task] button on the bottom right.
   • To save a draft and reload later, click the [Save draft] button.

Planning data

The planning data area shows all data of the model's ongoing thought process. You can check the thought process of the planning model that plans and executes the skills to use and their order when running user queries. The planning data area consists of 3 steps.

Step 1: planning step

This is the step to decide which skill to use.

Step 2: skill execution step

This is the step to call the skills. You cannot edit any of the fields.

Step 3: final response step

This is the step to decide the response to deliver to the user. Click the [Complete task] button to save all and finish creating the scenario.

Single skill data

The skill data area shows the skill data you want to call. The planning data area determines which skill to use for the user query, while the skill data area shows detailed data about the skill to be called and allows you to edit the data directly. The skill data area consists of 3 steps.

Step 1: search API

The first step for single skill data is to query an API.

Step 2: call API

The second step for single skill data is to call an API. The second step for skill data is to call an API. You can click each field to edit the value. When you edit the value of a field, the [Apply] button is activated. When you click the [Apply] button, the skill data is rerun, and the edited result is applied to the planning data area as well.

The following screen shows skill data that uses the GET method in the API call step.

When using the POST method, the field for entering the URL and parameters is activated in the action input area.
The following screen shows skill data that uses the POST method in the API call step.

Step 3: skill response

This is the step to decide the response to be delivered to the user after querying and calling the API.

Scenario configuration method

Depending on the number of skills used, you can configure either a single skill scenario or a multi skill scenario.

Single skill scenario

A single skill scenario is a scenario that uses just 1 skill. The scenario sequence is as follows:

Scenario sequence

1. User query: recommend travel destinations in Busan
2. Planning data Step 1: plan skill to use
3. Planning data Step 2: execute skill
   1. Single skill data Step 2-1: search API for travel destinations
   2. Single skill data Step 2-2: check API parameters for travel destinations, and call by using Busan as a keyword
   3. Skill data Step 2-3: return the result about travel destinations in Busan
4. Planning data Step 3: provide final response to user

Scenario description

In response to the user query "Recommend travel destinations in Busan," the model determines which skill to use. If the model determines that it needs to call the travel destination search API, it queries the API, checks the necessary parameters, and then calls the required information. Finally, the model decides in what form the result should be delivered to the user and then presents the response.

Multi skill scenario

A multi skill scenario is a scenario that uses 2 skills. The scenario sequence is as follows:

Scenario sequence

1. User query: recommend restaurants and car rentals near Gangnam Station
2. Planning data Step 1: plan skill to use
3. Planning data Step 2: execute skill
   1. Single skill data Step 2-1: search API for restaurants near Gangnam Station
   2. Single skill data Step 2-2: call API for restaurants near Gangnam Station
   3. Skill data Step 2-3: return the result about restaurants near Gangnam Station
4. Planning data Step 3: execute skill
   1. Single skill data Step 3-1: search API for car rentals
   2. Single skill data Step 3-2: call API for car rentals
   3. Skill data Step 3-3: return the result about car rentals
5. Planning data Step 4: provide final response to user

Scenario description
The model determines which skill to use in response to the user query that goes "Recommend restaurants and car rentals near Gangnam Station." Since search is needed for both restaurants and car rentals, the model determines that it needs to call 2 APIs. In the second step, the API that searches for popular restaurants near Gangnam Station is queried and called, and in the third step, the API that searches for car rentals is queried and called. Finally, the model decides in what form the information should be delivered to the user and then presents the response.

Load scenario

You can load, review, and edit scenarios you previously worked on.
To load a scenario, follow these steps:

1. On the NAVER Cloud Platform console, click Services > AI Services > CLOVA Studio in order.
2. Click the My Product menu, and then click the [Go to CLOVA Studio] button.
3. On CLOVA Studio, click the Skill trainer menu.
4. Click the [Collect scenario] button at the top of the screen.
5. Click the [Load] button on the bottom left of the scenario collection screen.
   • You can see the list of scenarios you are working on.
6. Check the previous query list.
   • You can search for user queries from the top right of the Load window.
   • To view the model result, click the [View] button of the relevant data item.
   • To delete a user query, click View more > [Delete] on the right of the list.
   • The task status shows "completed" only when you click the [Complete task] button on the scenario task screen.
7. To edit the previous query task from the scenario collection screen, click the [Open] button.
   • The scenario collection screen appears, and you can continue to work on the previous queries.

Create call option

Through the call option, you can configure settings to fix specific values of header and body parameters in response to the operation ID of the API spec. To create a call option, follow these steps:

1. Click [Setting icon] of the call option on the left side of the screen during the Collect scenarios process.
   • The call option settings window appears.
2. Enter the call option in JSON type in the call value input area. Create it referring to the example format above.
   • For more information on call value input, see Detailed call option.
3. Once the call value input is completed, click the [Apply] button.
   • If it fails to be applied, an error message will be displayed. Edit the call value again.

Detailed call option

It is information on the fields that can be written in the call option.

Example of writing a call value

You can use the baseOperation field when you want to apply a call option to all APIs registered within the skill set. The following shows how to write it.

{
  "baseOperation": {
    "query": {
      "serviceKey": "value"
    }
  }
}

You can use the operations field when you want to apply a call option only to a specific Operation ID. The following shows how to write it.

{
  "operations": [
    {
      "operationId": "filterBakeryProducts",
      "query": {
        "serviceKey": "value"
      }
    }
  ]
}

It is also possible to apply the baseOperation field and the operations field together. The following shows how to write it.

{
  "baseOperation": {
    "header": {
      "baseHeader": "baseHeaderValue"
    },
    "query": {
      "baseQuery": "baseQueryValue"
    },
    "requestBody": {
      "baseBody": "baseBodyValue"
    }
  },
  "operations": [
    {
      "operationId": "testOperationId1",
      "header": {
        "header1": "headerValue1"
      },
      "query": {
        "query1": "queryValue1"
      },
      "requestBody": {
        "body1": "bodyValue1"
      }
    },
    {
      "operationId": "testOperationId2",
      "header": {
        "header2": "headerValue2"
      },
      "query": {
        "query2": "queryValue2"
      },
      "requestBody": {
        "body2": "bodyValue2"
      }
    }
  ]
}