Auto-resolve customer tickets using the Email AI Agent

The Email AI Agent uses generative AI to automatically respond to customer queries via email using knowledge base content and contextual understanding. It helps customer support teams reduce manual workloads, improve first-response times, and maintain high-quality service at scale.

The Email AI Agent generates relevant replies based on ticket context and customer language preferences, all without human intervention.

Email AI Agent's Core Operational Workflow

1. Email Reception & Classification
   • Inbound emails are analyzed by a large language model (LLM)-based spam classifier (98% accuracy)
   • Non-actionable messages (newsletters, OOO replies) are automatically filtered
   • Actionable tickets proceed to language detection
2. Language & Intent Analysis
   • Detects query language independently of account settings
   • Supports 23 languages including Japanese (JA), Thai (TH), and Vietnamese (VI)
   • Identifies primary intent (e.g., refund request, technical issue)
3. Knowledge Base Query
   • Searches portal-specific public solution articles
   • Considers up to 3 most relevant articles
   • Applies confidence scoring to validate match quality
4. Response Generation & Delivery
   • Generates concise, cited responses in the customer's language
   • Formats with proper paragraphs, bullet points, and branding
   • Embeds feedback widgets for continuous improvement
5. Ticket Resolution
   • Automatically closes tickets receiving positive feedback
   • If the answer is unhelpful, human agents can always intervene. Tickets remain with agents, giving them full control while reducing their workload.

Reimagine Customer Experience with the Email AI Agent

The Email AI Agent improves customer interactions by providing quick, smart, and organized responses while maintaining the empathy and reliability of human support.

When a customer sends an email query to the support address, the Email AI Agent acknowledges receipt of the query by assigning a support ticket ID. This provides instant reassurance that the request has been logged and is being looked into.

The AI Agent analyzes the customer’s query and provides a well-structured response. The reply includes:

• A direct summary answer tailored to the specific question.
• A set of linked reference articles that offer deeper context and related information. This gives the customer options to self-serve based on their interest level.

At the end of the email, customers are invited to quickly share feedback via two clear call-to-action buttons:

• Yes, Close my ticket or
  Not really.

This simple interaction lets users control the outcome: if they’re satisfied, the ticket is closed instantly; if not, the case remains open for human follow-up.

Use Cases & Implementation Examples

1. E-Commerce Order Management

High volume of return/refund inquiries—especially post sale season—leading to long resolution times and inconsistent agent replies.

Implementation:
 Configured the Email AI Agent to respond to emails containing:

• Keywords like "return", "refund", "exchange", "wrong item"
• Language detection for multilingual template coverage
• Template-based summaries referencing official return and refund articles

Sample Workflow:
Customer emails: “Hello, I received the wrong item. Can I return it and get a refund?”
The Email AI Agent:

• Detects intent through keywords like “wrong item” and “refund”
• Identifies English as the language
• Summarizes content from the article “How to Return or Exchange a Product” and “Refund Timelines”
• Responds:
   "We're sorry you received the wrong item. You can return it within 30 days of delivery, as per our Return Policy. Once we receive the returned product, refunds are processed within 5–7 business days. For step-by-step instructions, please visit our Returns Help Center."

Outcome:

• Refund-related inquiries answered automatically with article summaries
• Reduced average handling time for human agents on transactional tickets
• Increased customer satisfaction by providing clear next steps instantly
  
  

2. SaaS Technical Support

24/7 coverage for a global customer base.

Implementation:

1. Created knowledge articles for:
   • Common error codes
   • API troubleshooting
   • Billing FAQs
2. Set escalation rules for:
   • Negative sentiment
   • "Critical" priority tickets
   • Mentions of "cancel account"

Sample Workflow:

Customer emails: "I can't authenticate with the new OAuth 2.0 implementation. Getting 'invalid_grant' errors when trying to generate tokens."

The Email AI Agent:

• Identifies authentication error pattern
• Sends: ‘For OAuth 2.0 'invalid_grant' errors:
  1. Verify your authorization code:
     • Was used within 5 minutes of generation
     • Hasn't been previously consumed
     • Matches your client ID/secret
  2. Check our OAuth migration guide [KB-587]
  3. If still failing, reply with:
     • Exact timestamp of attempt
     • SDK version used“

Auto-creates a support ticket "Auth Failure" tag if customer replies "still not working"

Outcome:

• Tier-1 issues resolved without human intervention
• Faster response time for critical outages

Prerequisites and Key Considerations before configuring Email AI Agent

Note: The Email AI Agent does not currently support responding to ongoing email threads — only initial ticket creation.

Components of Configuring Email AI Agent

Once you've accessed the Email AI Agent setup area, configuration involves five core components.

1. Trigger rules
2. Language configuration
3. Template editing
4. Advanced settings
5. Saving and enabling

1. Trigger Rules

You can define when the Email AI Agent should activate here. Trigger rules determine the criteria under which a ticket qualifies for automated replies. Trigger rules can utilize all available ticket, contact, and company properties - including system fields, custom fields, and behavioral attributes - to precisely determine when the AI agent should respond, ensuring it only engages with appropriate queries while avoiding irrelevant scenarios.

2. Language Configuration

Specify the languages your Email AI agent will operate in based on your customer base and available knowledge base content. The language selection dropdown displays only languages that meet all three criteria:

• enabled in your account settings,
• supported by the Email AI agent, and
• available across your portals.

You can enable or disable individual languages from this filtered list, and must define a primary language for fallback scenarios. This dual-layer validation ensures the Email AI agent only generates replies in languages that are both technically supported and properly configured in your knowledge base, maintaining linguistic appropriateness and contextual accuracy. The system automatically synchronizes these language settings across all connected portals while respecting each portal's individual content availability.

3. Template Editing

You can customise the AI-generated email templates for each supported language. Templates define the structure of the response, including how article suggestions and feedback components are embedded along with summarised answers. You can fine-tune tone, formatting, and placeholder positioning to align with your brand’s voice.

4. Advanced Settings (Limits and Controls)

Apply operational safeguards like daily ticket response limits to control how many tickets the Email AI Agent handles. These limits help manage exposure during rollout, prevent over-automation, and give you time to evaluate performance.

5. Saving & Enabling

Finalize your configuration by saving changes and enabling the Email AI Agent. This validates that all necessary components (e.g., language templates, triggers) are in place. You can also choose to save without enabling, allowing for staged rollouts or peer reviews.

Analytics & Reporting

The Analytics module provides a comprehensive view of how the Email AI Agent is performing across your support operations. These insights help admins evaluate effectiveness, troubleshoot issues, and make informed decisions about language support, template tuning, and knowledge base quality. The Freddy Email AI Agent curated report has the following widgets:

1. Tickets Answered By The Email AI Agent
2. Email AI Agent response rate
3. Sessions consumed
4. Ticket Deflection Trend
5. Sessions consumption trend
6. Email AI Agent Response Rate Breakdown
7. Email AI Agent Overall Helpfulness Feedback

Learn more about Email AI Agent Performance report.

See More

• Advanced Configuration Guide
  
• Related Features:
  • Sentiment Analysis for Ticket Prioritization
  • Freddy Article Suggestions
  • Auto-Triage

If you signed up before Aug 9, 2021, please click Previous plans to view your applicable plans.