/ /
Using Agent Studio (EAP)
Updated 31 seconds ago

Overview

Note: This feature is part of our Early Access Program, which allows select customers to try upcoming beta features that are clearly labeled in the product. As part of this program, you can explore new capabilities, provide feedback, and help shape their development. To join the Early Access Program, contact your Customer Success Manager (CSM) or Account Manager (AM) to request access.

The Agent Studio can perform complex tasks, giving teams control and creativity in how they use Responsive AI. It’s an intuitive, user-directed, no-code environment where anyone can design intelligent agents that fit their unique workflows and goals.

Users can build strategic, intelligent agents for higher-value use cases such as market research, competitive analysis, and content governance. It provides AI-powered insight generation and analysis, giving teams more flexibility to create agents.

  • Agents created in Agent Studio automatically retain preferences across different sessions. This means that once you create an agent with Always respond with bullets or Always be concise, those preferences are automatically applied to the next agents you create.

  • If you change your mind, just tell the agent to stop using that specific format and it resets its instructions.

Using Agent Studio

Select Agent Studio from the 9-dot menu. Any existing agents appear on the page. You can click one to run it, search for an existing agent, or design an agent by entering what you want to create.

When you edit an agent, all users that can see it receive an email notification whenever:

  • An unpublished agent is republished.

  • When a user or group is removed from the user list after it is published or republished. If a group is removed, all members of that group will be notified.

  • If an agent’s visibility changes to Private.
    Note: If the visibility changes from specific users to Public, no notifications are sent.

  • If an agent’s status changes to Draft (the owner will not be notified).

Defining your agent

  1. Click inside the dialog box and describe what you want your agent to do, then press Enter.
    image.png

  2. Enter a name for your agent. Be sure to choose something that makes your agent easy to identify.
    image.png

  3. Describe what the agent should do in natural language. This information is used to pre-populate the rest of the agent.

  4. (Optional) Upload an image to help identify your agent.

  5. Click Next.
    Note: You can click Exit at any time to stop creating your agent; it won' be deleted.

  6. Describe what you want the agent to do, and how you want it to behave. You can include goals and samples if necessary.
    image.png

  7. Do the following as necessary:
    image.png

    • Click Refine Prompt to refine the default instructions.

    • Click Undo to revert to the previous instructions.

  8. Click Next. Based on the instructions, Responsive AI automatically populates the agent details, instructions and input fields.

Defining prompts to users

  1. Specify how the agent will get information from the users. 
    image.png

    • Agents automatically retain your stylistic and formatting preferences across your sessions. For example, if you instruct an agent to always answer in paragraphs that rule is automatically applied to all future interactions with that specific agent.

    • This is designed to make conversations with the agents you build in Agent Studio more seamless. Every agent you build listens to your preferences and remembers them. Whether you prefer detailed paragraphs, bullet points or a specific tone, you only need to state it once for that agent and it's automatically applied to all your future chats. 

  2. (Optional) Click Add Field to add more fields:

    • Enter the name of the field you want to add

    • Click the toggle if you want to make the field required

    • Enter any help text you want the user to see

    • Specify the field type:

      • File Upload: Allows the user to upload a file to be acted upon. You can have up to two fields with the File Upload type in your agent.

      • Text: Such as names, usernames, and URLs (up to 130 characters).

      • Single Select: The agent provides a list of options and users can select one of them. The list can have up to 10 options.
        image.png

      • Multi Select: The agent provides a list of options and users can select more than one of them. The list can have up to 10 options.
        image.png

      • Date Picker: The agent displays calendar and allows users to choose a date or date range.
        image.png

  3. Enter the instructions for the agent about how to use the information.
    image.png

    • Click the Trash icon to delete a field.

    • Use the handles to drag and drop the fields to rearrange them.

  4. Click Next. When the user runs the agent, they're prompted to add information in those fields or upload files like in the example below.
    image.png

Connecting your knowledge sources

Specify where the agent should go to get information to perform the tasks. This can be your own Content Library or an external source like a SharePoint site or a Google drive location. After connecting to those sources, the agent you're creating can access them to generate answers and take action.

  1. Specify where your agent should go to get the information used to perform tasks:
    image.png

    • Turn on the Responsive toggle if you want the agent to look in your Content Library.
      image.png

      1. Specify the collections, tags, and response headers you want to use.

      2. Click + Add Filter to add more filters.

        • Any: Content will be used if it has any of the tags or collections

        • All: Content will be used if it has all the specified tags and collections

        • Exclude: Content will not be used if it has the specified tags or collections

    • Click Add New Source to add external data sources (Google, SharePoint or websites).
      Note: Your Google Drive and/or SharePoint accounts must be connected to Responsive before using this feature. See Enabling cloud storage integrations and Microsoft SharePoint integration guide for more information.

      • For Google Drive:

        1. Click Google Drive and log in to your account, then select the folder you want to add.

          • There is no limit to how many folders you can add but, by default, only individual folders 1 GB or smaller can be crawled. If your folders are larger, contact the support team for assistance.

          • Crawling supports the following document formats: .pdf, .docx, .doc, .xlsx, .xls, .xlsm, .pptx, .ppt.

        2. Do the following as needed:

          • Click Add more folders to add additional folders.

          • Specify the collection you want the folder added to.

          • Click the Remove icon to remove any unwanted folders.

        3. Click Save.

      • For SharePoint:

        1. Click SharePoint, select a site from the drop-down list, then click Add Folders.
          Note: The folder-level permissions set in SharePoint carry over to Responsive.

        2. Click the folder you want to add and click Select.

          • There is no limit to how many folders you can add but, by default, only individual folders 1 GB or smaller can be crawled. If your folders are larger, contact the support team for assistance.

          • Crawling supports the following document formats: .pdf, .docx, .doc, .xlsx, .xls, .xlsm, .pptx, .ppt.

        3. Do the following as needed:

          • Click Add more folders to add additional folders.

          • Specify the collection you want the folder added to.

          • Click the Remove icon to remove any unwanted folders.

        4. Click Save.

      • For websites:

        1. Click Website and do one of the following:

          • Click All Public websites to use the entire internet as a knowledge source.

          • Click Selected Websites and enter the websites URL.

            • You can add up to 10 domain URLs.

            • Make sure to enter help.responsive.io as one of the URLs if you want your agent to search the Responsive Help Center.

        2. Click Add.

  2. Click Next. If the agent was published before, click Republish to begin indexing the new knowledge source(s).

Launching your agent

  1. Review your agent and update the visibility settings if needed.

    • The visibility is set to private by default. When you're ready to use the agent, you can change it to Public or Specific users.

    • If you change it to Specific users, assign the users and click Save.
      Note: Only the owner of the agent can share it. Users receive a notification whenever a new agent is made available to them and changed.

  2. Click Publish.
    image.png

  3. Click Launch Agent to make it available for use.
    Note: You can click Unpublish if you want to make more changes.
    image.png

Invoking an agent 

Note: Agents are invoked from the Ask menu, not from Agent Studio. Ask is the primary interface where users will interact with agents. 

  1. Click Agents or Browse Agents to see the available agents.
    image.png

    • The agents you have will be displayed, with the most recent at the top.
      image.png

    • Each one is labelled so you can tell which ones were provided by Responsive (Responsive Agent) and which ones you created (Agent Studio).

    • You can search to find a specific agent.

  2. Click the agent to invoke it in a separate browser tab.

Adding Canvas results to section descriptions in a project

Note: This feature is available for agents you create using Agent Studio; you cannot use it for agents provided by Responsive.

  1. Run an interaction with the agent so the generated response is in the Canvas.

  2. Click the three-dot Menu and select Save to Projects.
    image.png

  3. Click Save as Section and select the project, then select a section from the list. Click New Section to create a new section and add the content as the description.
    image.png

    • If the section already has a description, you'll be prompted to specify if the generated content should be appended to the description or if the description should be rewritten to include the content.
      image.png

    • If it does not, the generated results become the description of the newly created section.

Adding Canvas results as documents in a project

Note: This feature is available for agents you create using Agent Studio; you cannot use it for agents provided by Responsive.

  1. Run an interaction with the agent so the generated response is in the Canvas.

  2. Click the three-dot Menu and select Save to Projects.
    image.png

  3. Select Save as a Project Document, then click Save.

  4. Specify the project, document type (currently only Word is supported), and the name you want the document to have.
    image.png

  5. Click Save.

  6. Do the following as needed:

    • Click View in the success message to see a preview of the document.
      image.png

    • Download the document while the preview is being generated.
      image.png

    • You can view the generated document, download, and print it.
      image.png

    • Go to the project to see the generated content as a document for the project you specified.
      image.png

Was this article helpful?
Subscribe to receive updates on this article