Promptbook

Advanced integration of an AI agent into the ticketing system for support automation and intelligent task management.

Project Goal

Implementation of an advanced AI agent integration into the internal ticketing system, which automatically reacts to new tickets and performs initial triage plus action steps.

• Automatic reaction to new tickets
• Initial triage and action steps
• Faster request handling
• Scalable support system

Process Workflow

1. Ticket creation

The user creates a new ticket via any of the available channels.

2. System processing

The system assigns an ID and basic parameters; the ticket is placed into the queue.

3. AI analysis

The AI agent performs analysis and decides on the next steps.

4. Automated actions

The agent performs the appropriate actions based on the type and content of the ticket.

5. Workflow continuation

The ticket continues through the standard process or is automatically closed.

Input Channels for Tickets

The system supports several ways to create a new request:

• E-mail — automatic processing of incoming e-mails.
• Admin interface — direct ticket creation in the administration.
• Contact form — form on clients' websites.
• Support web — dedicated site for technical support.

AI Integration — Technical View

Input data

The AI agent receives a JSON payload with the complete ticket context, including:

• Ticket type and category
• Labels and priority
• Ticket text and problem description
• Metadata and parameters
• History of similar tickets

Decision process

Based on data analysis the AI agent decides on next steps:

• Classification of the problem type
• Determination of priority and urgency
• Selection of an appropriate assignee
• Proposal of initial steps
• Automated actions

Possible AI agent actions

• Add a public or internal comment
• Change ticket state (open, in progress, closed)
• Assign labels and categories
• Change priority based on severity
• Assign a responsible person
• Add documentation links
• Call the internal BizKitHub API
• Send notification e-mails

Data Model

Main entities of the system:

Core entities

• issue — main ticket
• project — project / organization
• issueStatus — ticket states
• SLA — service level agreement

Supporting entities

• user — system users
• comment — comments
• attachment — attachments
• aiAction — AI actions

Note: A detailed database table design has been prepared and the structure continues to evolve as implementation progresses.

Benefits

• Faster handling — automated processing reduces first-response time.
• Scalable support — AI can process a large volume of tickets in parallel.
• Consistent quality — standardized procedures ensure uniform support quality.
• Intelligent triage — automatic problem-type recognition and assignment to the correct handler.

Long-Term Vision

1. Expansion to internal queries

The AI will handle not only external tickets but also queries from internal employees, using full-text search across the whole organization.

2. Contextual recommendations

The AI will be able to surface relevant context and recommendations based on historical data and the organization's knowledge base.

3. Scalability and auditability

The system will be fully scalable, auditable, and customizable for different clients and SLA requirements.

Technical Specifications

API integration

• REST API communication
• JSON payload format
• Authorized API calls
• Real-time webhook notifications
• Audit logs for all actions

Security and scaling

• Isolated conversations
• Request-response pattern
• Protection of sensitive data
• Horizontal scaling
• Monitoring and alerting

Implementation Status

Promptbook AI integration is fully implemented and actively used in the BizKitHub production environment.

Completed

• Technical integration
• AI decision logic
• Automated actions
• Monitoring system

In operation

• Ticket processing
• Automatic classification
• Intelligent routing
• Performance optimization

Future

• Feature expansion
• New AI models
• Advanced analytics
• Multi-tenant support