Zia actions in Workflows: Field Extraction and Field Prediction are opened in a phased manner and will be available only for selected users. We will soon roll it for all users. These features are supported in Enterprise edition.
Importance of AI in Workflows
Business processes are the foundation of any organization, and workflows are the pillars that uphold them and guarantee operational continuity at all times. Workflows streamline business processes by reducing redundancy and optimizing output using cost-effective measures, which ultimately improve process and team efficiency.
Workflows orchestrate a process by automating manual actions at the relevant stages that significantly reduce human intervention, freeing up people to invest time in various critical activities that require their insights.
In synergy with Artificial Intelligence, workflows operate smarter as they analyze the data to extract information, predict values, send auto-replies, and generate contextual text from ticket replies. Overall, this kind of backstage automation creates a more efficient and responsive support experience for the customers and streamlines operational activities by automating routine steps.
Zia actions in Workflow
Tickets are a goldmine of information. They consist of customers' details, such as email address, phone number, best time to reach, subscription plan, account type, along with a detailed description of their issue, service, or product of complaint.
A ticket is classified, triaged, routed, and prioritized based on these details to ensure quick appropriate resolution which is critical in terms of SLA adherence.
Zia can intercept useful details in newly created tickets and accelerate the automation flow.
Below is a typical customer service scenario where a customer has reached out to the XYZ Products Ltd. support team for an enquiry on their order.
Alex sends an email about the return of a product that was wrongly purchased a week ago. Since XYZ's website doesn't have the return policy marked-up clearly, Alex sends an email to the support team (
xyz@support.com) to find out whether the product will be picked up by the company or he should courier it to the company's address.
In the email, Alex shares the transaction details such as "Date of purchase", "Mode of payment", and "Product condition". He also wants to know when the paid amount will be re-credited to his account.
As per the ticket criteria defined in the workflow configuration, the ticket enters a specific workflow. From here, Zia will analyze the ticket's content and take the following actions:
Action type | Action | Result |
Data extraction | Extract customer information and case details such as email address, contact number, best time to contact, and issue type. | - Autofill the extracted details into the relevant fields in the ticket layout.
- The ticket is auto-classified as per the issue type. If a workflow for a specific issue type is created, it will be automatically applied on the ticket.
|
|
Field prediction | Analyzes the content to identify details such as "Product type", "Mode of payment", "Product condition" etc. | Autofill the extracted details into the relevant fields in the ticket layout.
|
Auto triage using field prediction inputs | Predict the case severity (eg., Critical, High, Medium) and type of support service needed (eg., L3, L2).
| - Autofill the predicted severity and support level fields.
- The ticket is auto-triaged and assigned to the respective team or agent based on the populated field values. |
Generate content | Summarize and highlight the key points discussed in the email and the threads. | - Agents will get a gist of the issue by reading the summary. - Summary is added as a private comment for other stakeholders to refer, if they are involved at a later stage of resolution. |
Email auto-reply | Generate a suitable reply with article links for the customer to read and refer to. | - An automatic reply is drafted and sent to the customer.
- The article links, such as Return policy, Purchase guidelines, Common FAQs, help the customer understand and troubleshoot the problem on their own. |
Outcome of the above orchestration
- The speed at which things are done is significantly higher.
- Accuracy of the resultant actions is better.
- FRT and ticket reopening rates are considerably low.
- Consistent replies that follow business guidelines.
- Agent efficiency is improved as much of the routine work is taken care of by Zia.
- Improved CSAT with customers getting quick, accurate replies instantly.
- Better business outcome with processes becoming more agile and efficient.
Setting up Zia actions in Workflows
Points to remember
- Administrators can configure Zia actions in the workflows.
- Zia action is available ONLY in Enterprise edition.
- Zia actions can be configured ONLY for the Tickets module.
- Limits: These limits are only applicable to Zia automation actions and do not affect Zia native features or AI agents.
- A workflow can have a maximum of 1 type of each Zia action at a time. That is, 1 Workflow rule can have all 4 Zia actions.
- An action can be used only once in a rule.
- A maximum of 1000 executions/org/day will take place (includes all Zia actions).
- Generate and Auto email reply are released in a phased manner. Kindly reach out to the Desk support team at: support@zohodesk.com to enable the feature for your org.
Availability
- Users who have the permission for Help Desk Automation can setup Zia actions in the Workflows.
- Zia actions is supported in Enterprise edition.
Prerequisites
- Zia actions are governed by generative AI services, therefore Generative AI must be enabled before setting up the Workflow from the Setup > Zia > Generative AI. Read more Setting up Zia GenAI services
- Auto email reply: Both Generative AI services and Answer Bot must be enabled for the Auto email reply action to operate. Department based Answer bot must be created and trained prior to the setup. Read more Understanding Answer Bot
- Email replies and content generation is supported only in English language.
Zia can extract values such as email, phone number, due date, case type, product code, order number, purchase date etc from the ticket conversation and autopopulate them in the ticket. This saves time and considerably reduces the chances of missing out on important details and inadvertent errors while entering values manually.
Supported fields: Zia will identify ONLY the below mentioned fields from the email text, extract them, and auto-fill them in the respective fields. It is recommended to verify the values to ensure complete accuracy.
- Number, Decimal, Percent
- Currency, Email, Phone, URL
- Date and DateTime
Limitations:
- A maximum of 5 field values can be extracted/rule.
Some industries that can benefit from this are:
Universities and education verticals: A student sends an email seeking assistance for an issue with accessing study materials for a course.
- Zia intercepts the email, identifies the fields present in the layout, and maps them with the details mentioned in the email (eg., Exam code, Student ID, Contact number, Registration date, Admission follow-up date etc.).
- Extracts the field values from the email and auto-populates them.

Insurance and mortgage: A policy holder raises a ticket for an issue with the repayment of mortgage debt. In addition, the person wants a detailed breakdown of the overbearing associated costs.
- Zia analyzes the email and extracts the information needed to move the ticket to the next stage.
- It identifies policy holders' name, contact details, policy name and number, issue type, and description of the issue.
- Auto-fills the values in the respective fields.
- The ticket is automatically assigned to the Debt handling team in the mortgage department.
To configure Field Extraction in Workflows
- Go to Setup > Automation > Workflows.
- Click Create New Rule.
- Select Tickets Module.
- Enter Rule Name, Description, and click Next.
- Select when to Execute the rule.
- Select a Criteria, if needed.
- Under Actions, select Field extraction.

- In the Field Extraction page, do the following:
- In Text to analyze, select the fields from the placeholder in the right panel that Zia should analyze and extract values from.
- In Field and sample format, select the fields (email, phone number etc.) from the drop-down and provide a sample format (mm/dd/yyyy; 123-0000-999). Providing sample format is optional.
- Click Save.

Field prediction
Zia Field prediction and Zia Workflow action Field prediction are distinct features. The Zia workflow action based prediction does not require training.
Zia can analyze the email content, thread replies, subject, description, and several other fields available in the layout to predict values for specific fields. It evaluates the content and automatically maps the values to the most relevant field.
For example, it can analyze the description and the thread replies in the ticket to predict when the student would opt for an admission and fill the correct value in the field Admission on: "Autumn 2025 batch". Likewise, it can predict the VISA status and update the field to"Awaited" or "Delayed". If an email notification is sent to the VISA handling team, they can reach out and provide the necessary assistance.
Supported fields: Zia can predict values for the following fields:
- Picklist
- Colored Picklist
- Multi-select Picklist
What happens when both Zia Field Prediction and Field Prediction action are configured?
The value predicted by Workflow prediction will overwrite Zia's prediction.
For example, when a ticket is received if Zia predicts the 'Problem Type' as 'Manufacturing Defect' and later the customer sends an email with more details and field prediction action analyzes the content and intercepts the 'Problem Type' as 'Faulty Device' the original predicted value will be replaced with 'Faulty Device'.
To configure Field prediction as an action in Workflow
- Go to Setup > Automation > Workflows.
- Create New Rule.
- Select Tickets Module.
- Enter Rule Name, Description, and click Next.
- Select when to Execute the rule.
- Select a Criteria, if needed.
- Under Actions, select Field Prediction.
- If the Field prediction page, do the following:
- In Text to analyze, select the fields from the placeholder that Zia should analyze to predict the field values.
- Select the Field that Zia should predict a value for.
- Select the values that Zia should predict.
- Click Save.

Generate content
This feature is released in a phased manner; it will soon be available for all users across DC. In the meantime, if you want to try Content Generation in Workflows kindly reach out to
support@zohodesk.com Customer emails can be long and sometimes confusing. Agents may find it difficult to comprehend the exact problem; they may even overlook important details. Such missteps can be avoided if the email content can be summarized into a meaningful gist and added to the ticket as a comment or a field.
Zia can summarize the customer requirement and highlight the main points discussed in the email. It presents it in a concise, easy-to-consume text for the agents to take quick appropriate actions. The summary is auto- populated in the private comment, public comment or in a multi-line text field as configured in the workflow.
It can analyze fields such as recent thread content, ticket description, ticket status, priority level, and subject to understand the context and generate a summary. It also takes into consideration the prompt instructions provided by the user to generate a summary.
For example, if a customer sends an email for onboarding with an on-site implementation request and enquiring about a long-term facility to provide a continuous training certification to its employees, Zia can analyze the thread content, ticket subject, description, and prompt, such as: "Summarize the ticket conversation, highlight the list of requirements, and mention the urgency of the request" to generate a summary which can be added as a private comment for the concerned stakeholders to refer.
The presales, customer onboarding, solutions, and implementation teams need not read the customer's email to understand the context and requirement; they can refer to the summarized content available in the ticket comment to take the further actions. It saves time and significantly reduces the chances of inadvertent overlooks or misunderstandings that are common when multiple stakeholders are involved.
Points to remember
- Generative AI services must be enabled before configuring content generation in Workflows.
- Content will be generated ONLY in English language.
- For GenAI services using ChatGPT there will be token charges as per the usage. For more information on this read: ChatGPT as GenAI services in Desk
To configure Generate content as an action in Workflow
- Go to Setup > Automation > Workflows.
- Create New Rule.
- Select the Module.
- Enter Rule Name, Description, and click Next.
- Select when to Execute the rule.
For Zia Email Reply, select Create. - Select a Criteria, if needed.
- Under Actions, select Generate content.
- If the Generate content page, do the following:
- In Text to analyze, select the fields from the placeholder that Zia should analyze to generate a summary.
- In Action to perform, select Private comment, Public comment or Field update.
- In Prompt, enter your instruction or suggestion that Zia should consider while summarizing.

- Click Save.
Auto email reply
This feature is released in a phased manner; it will soon be available for all users across DC. In the meantime, if you want to try Content Generation in Workflows kindly reach out to
support@zohodesk.com Points to remember
- Auto email reply is supported only for the tickets created through the Email channel. It can be executed only upon "Creation" of a ticket.
- The reply will be generated and sent ONLY for the First response.
- Answer Bot must be configured and trained using the KB articles before setting up Zia email replies. This is important as Zia can draft a response and provide suitable knowledge base articles in the email only if Answer Bot is trained to identify and analyze similar queries. Read more about AI-based self-service through Answer Bot
- Ensure that Answer Bot is configured in the same department where the Workflow is configured.
- The auto-replies are generated ONLY in English.
Auto reply will not generate a reply if:
- Ticket is not received by Email channel.
- There is no incoming email thread.
- The email content doesn't have much detail or has unclear language.
- An outbound email is sent in the meantime.
Using workflow, automatic emails replies can be sent to customers' tickets with answers that are generated by the Answer Bot along with related article links from the knowledge base.
For example, enquiries about product installation, issues with setting up accounts, and password retrievals can be auto-resolved using Answer Bot's information retrieval and Zia's email composing capabilities.
This significantly reduces the number of tickets that agents handle manually. Smart routing and auto-resolving of tickets can improve agent productivity and overall process efficiency. A wide-range of technical problems, IT related issues, and maintenance glitches can be categorized and easily resolved by AI using the company's troubleshooting tips, user manuals, or best practice guides that are cost-effective ways to improve support operations across the company.
The support teams can instead focus on resolving more complex issues, providing periodic training, or conducting paid educational programs for its customers.
The configuration also allows the admins to check the success rate of email delivery. They can update certain fields when a reply is sent or not sent, using which the failure rate can be tracked using a custom view "AI-based Auto Reply Dashboard". For example,
Auto reply will not generate a reply if:
- If Zia sends a reply, then update 'Auto response sent" as "Yes".
- If Zia doesn't send a reply, then update "Auto response sent as "Failed".
To configure Auto email reply as an action in Workflow
- Go to Setup > Automation > Workflows.
- Create New Rule.
- Select the Module.
- Enter Rule Name, Description, and click Next.
- Select when to Execute the rule.
For Email Auto Reply, select Create. - Select a Criteria, if needed.
- Under Actions, select Auto email reply.
- If the Auto Email Reply page, do the following:
- Ticket ID is auto-populated and cannot be modified.
- In the User Reply Address, select the agent's email address.
- In From Address, select the email address.
- In To Address, select an email address field (eg. contact's email, email field in the ticket) from the placeholder.
- In the Email Reply Template, modify the provided response if needed.
- In After Zia sends a reply, select a field and its corresponding value.
- In If Zia cannot send a reply, select a field and its corresponding value.
- Click Save.
