FAQs: AI-powered Workflow Automation with Zia Actions

FAQs: AI-powered Workflow Automation with Zia Actions

What are the key benefits of using Zia Actions in workflows?
  1. Faster ticket handling - Automatically extracts details like order number, purchase date, or product type from a customer’s email and fills them into the ticket. For example, if a customer emails, “I bought a phone on Sept 15 using PayPal.” Zia extracts purchase date and payment mode and updates the fields instantly with no agent data entry needed.
  2. Consistent first responses - Generates auto-replies with knowledge base articles so customers get quick, standardized answers. For example, if a customer asks, “How do I reset my password?” Zia replies instantly with a step-by-step guide and a link to the FAQ article or the available knowledge base content.
  3. Better agent productivity - Summarizes long customer emails into a short gist for agents, so they don’t have to read through every detail. For example: Instead of reading a 5-paragraph email, the agent sees: “Customer requesting refund for Order #2345, paid via credit card, product unused.” This saves time and reduces errors.

What types of Zia Actions can be configured in workflows?
Zia Actions are available ONLY for the tickets module. There are four Zia Actions user can set up in workflows:
  1. Field Extraction - Extract key details (like email, phone, order number) from tickets and auto-fill them into the tickets layout.
  2. Field Prediction - Predicts values for picklist and multi-select picklist fields (e.g., severity, product type, issue category) based on ticket content.
  3. Generate Content - Creates contextual AI-generated text such as summaries, highlights of key points, or concise explanations based on the ticket conversation for agents.
  4. Auto Email Reply - Sends AI-generated replies to customers with knowledge base article links. Only first responses are sent by Zia.
How do I enable Zia Actions?
Before using Zia Actions in workflows, here is what needed to be enabled:
  1. Generative AI - This needs to be turned on in Setup > Zia > Generative AI. Enable the Zia actions toggle.

  2. Answer Bot - For Auto Email Reply, configure and train the Answer Bot with the Knowledge Base articles. It should be set up in the same department where the workflow is created. Learn how to train Answer Bot.


How to set up Zia Action in a workflow?
Zia Actions allow workflows to use AI to extract data, predict fields, summarize content, and send auto-replies. Any user with Automation permission can set up Zia Actions in workflows.
To set up a Zia Action in a workflow
  1. Enable the following prerequisites:
    1. Ensure Zia Actions in Generative AI services are enabled (Setup > Zia > Generative AI).
    2. For Auto Email Reply, configure and train the Answer Bot in the relevant department.
  2. Navigate to Setup > Automation > Workflows.
  3. Click Create New Rule.
  4. Choose the Tickets module.
  5. Enter a Rule Name and Description, then click Next.
  6. Select when to execute the rule (e.g., on ticket creation or update).
  7. Add any necessary criteria to apply the action only to specific tickets.
  8. Under Actions, select the desired Zia Action:
    1. Field Extraction
      1. In Text to analyze, select the fields from the placeholder that Zia should analyze and extract values from.
      2. In Field and sample format, choose the field type (e.g., email, phone number) from the drop-down. Providing a sample value is optional.
    2. Field Prediction
      1. In Text to analyze, choose the fields Zia will use to predict values.
      2. Select the field to predict.
      3. Choose the possible values Zia can assign.
    3. Generate Content
      1. In Text to analyze, choose the fields Zia will use to generate the summary.
      2. In Action to perform, select Private comment, Public comment, or Field update.
      3. In Prompt, enter instructions or guidance for Zia to follow while summarizing.
    4. Auto Email Reply
      1. From Address: Choose the sending email.
      2. To Address: Select the customer’s email field.
      3. Prompt: Type in the prompt to help Zia to generate response.
      4. Email Reply Template: Edit the draft if needed.
      5. After Zia sends a reply: Choose a field and value to update.
      6. If Zia cannot send a reply: Choose a fallback field and value.
  9. Click Save to activate the workflow with the Zia Action.
How does auto email reply work?
In Auto Email Reply, Zia automatically drafts and sends the first response to a customer whenever a ticket is created via the email channel. Zia analyzes the content of the incoming email and, using the Answer Bot, drafts a suitable reply (based on the training using the Knowledge Base articles). The response may include helpful links or instructions to guide the customer in resolving their issue on their own. There are instances when Zia may be unsuccessful in sending a response:
  1. Unavailability of the information in the KB. 
  2. Lack of training in the respective area.
  3. Inability to generate a response due to content or training limitations.
  4. Zia was not continually trained to understand trending issues.
  5. Insufficient or unclear information in the customer’s email content.
  6. If an agent has already responded to the ticket, the auto email reply triggers only once, for the first response.
Support managers can monitor and track Zia's success rate by using the field update option. During configuration, admins can mention the field that should be auto-updated when a response is successfully sent or failed. A custom view can be created to get an exclusive view.

To configure Auto Email Reply, Generative AI and Answer Bot must be enabled and trained in the respective department.
  1. Navigate to Setup > Automation > Workflows
  2. Click Create New Rule.
  3. Select the Tickets module, enter a Rule Name and Description, then click Next.
  4. Set the rule to execute on Ticket Creation.
  5. Under Actions, select Auto Email Reply.
  6. Configure the email settings:
    1. From Address - Choose the sender email
    2. To Address - Select the customer’s email field
    3. Email Reply Template - Review or customize the reply
    4. Field updates - Track success or failure of the reply
  7. Click Save.
What kind of fields are predicted by Zia?
Zia can predict both standard and custom fields in tickets, as long as they are either picklist, colored picklist, and multi-select picklist.
Common examples can be:
  1. Category - Billing, Technical Issue, Refund
  2. Priority - High, Medium, Low
  3. Product or Service Type - Smartphone Model X, Premium Subscription, Home Insurance Plan
  4. Problem Description - Device not turning on, Payment failed, Policy renewal query
  5. Issue Type - Server Down, Network Connectivity, Software Installation, Password Reset, System Crash
Zia learns from historical ticket data to predict the most likely value for these fields whenever a new ticket is created or updated.
How is Field Prediction in Zia Action different from Zia Field Prediction?
  1. Zia Field Prediction - Zia Field Prediction uses a trained AI model that learns from your past data to predict field values automatically. It relies on historical patterns and accuracy thresholds, and updates prediction fields based on how confident the model is.
  2. Zia Action's Field Prediction - Zia Action Field Prediction works as part of workflow automation. Instead of using a trained model, it analyzes the content of tickets or emails in real time and instantly fills picklist fields. It can also overwrite the values predicted by normal Zia Field Prediction, making it ideal for immediate, content-based automation.
What kind of content does Zia generate?
Zia generates contextual texts based on the ticket conversation and user-defined prompts. The generated output can include:
  1. Summaries of long or complex customer emails
  2. Key highlights or bullet points from the conversation
  3. Contextual explanations or short descriptions
  4. Concise versions of multi-thread discussions
  5. Prompt-based text, such as listing requirements, noting urgency, or highlighting action points. Any kind of answer or text requested in the prompt, as long as it can be derived from the ticket content.
This content can be inserted into a private comment, public comment, or field update, based on how the workflow is configured.
To add the generated content within a workflow:
  1. In the workflow rule, under Actions, click +.
  2. Select Generate Content.
  3. In Text to analyze, choose the fields from the placeholder that Zia should review to generate the output.
  4. In Action to perform, choose whether the content should be added as a Private Comment, Public Comment, or Field Update.
  5. In Prompt, enter the instruction that Zia should follow while generating the content.