Skip to content
Znuny Logo Znuny Community Docs

Znuny AI: Ticket Classification, Assistants and Automation

Znuny AI means connecting external models or services to Znuny workflows. Znuny remains the system of record; the integration can analyze ticket text, suggest fields, or return routing decisions through controlled interfaces. This guide does not assume that Znuny ships a native AI suite or chatbot.

Practical Znuny AI use cases

Section titled “Practical Znuny AI use cases”
  • Ticket classification: suggest a queue, type, priority, or selected dynamic fields from subject and article text.
  • Information extraction: turn unstructured requests into fields that agents can review.
  • Agent assistance: summarize long conversations or draft a response for human approval.
  • Self-service: connect a separate assistant to a knowledge source and hand unresolved requests to Znuny.

Automation should start in suggestion mode. Measure accuracy, define confidence thresholds, and keep a manual fallback for ambiguous or high-impact tickets.

Automated ticket classification

Section titled “Automated ticket classification”

A classifier receives only the ticket fields it needs, returns a prediction, and passes the result to an integration layer. The integration validates allowed values before updating Znuny. OpenTicketAI for Znuny is one documented option for on-premise classification; a custom service can also use the Znuny REST API.

Before enabling automatic routing:

  1. define the queues and fields that may be changed;
  2. test with representative historical or synthetic tickets;
  3. set confidence thresholds and a fallback queue;
  4. log predictions, overrides, and processing errors;
  5. review quality regularly as ticket patterns change.

Chatbots and assistants

Section titled “Chatbots and assistants”

A Znuny chatbot is normally a separate application connected to the customer portal or an API. Depending on the implementation, it can answer approved FAQs, collect structured details, or create a ticket for an agent. Do not let generated answers bypass access controls, approval rules, or the documented handover process.

Privacy and operations

Section titled “Privacy and operations”

Ticket text may contain personal, contractual, or security-sensitive information. Minimize transmitted fields, document retention, restrict service credentials, and verify where inference and logs are stored. For regulated environments, an on-premise design can reduce external data transfer, but it still requires patching, monitoring, backups, and access controls.

Track classification precision, manual overrides, routing failures, and time to first response. For broader service reporting, see Znuny statistics, Power BI, and BI reporting.

Summary

Section titled “Summary”

Znuny can be extended with AI for ticket classification, extraction, assistants, and routing without presenting those integrations as built-in features. Start with a narrow workflow, keep agents in control, and expand only after measured results are reliable.

Discover OpenTicketAI for Znuny

Integrate AI-driven ticket classification and intelligent routing into Znuny — while retaining 100% on-premise data security.

Learn more about OpenTicketAI

Frequently asked questions

Does Znuny offer built-in AI capabilities or a native AI suite?

No, Znuny itself does not ship with a native, all-encompassing AI suite or a built-in chatbot. Instead, Znuny is designed to integrate with external AI models, services, or APIs. This approach allows organizations to leverage specialized AI capabilities for tasks like ticket classification, information extraction, or agent assistance, while Znuny remains the core system of record for ticket management. Integrations are typically handled through controlled interfaces, ensuring that AI-driven suggestions or actions are validated before updating Znuny.

Quellen / Sources

What are the main practical applications of integrating AI with Znuny?

Integrating external AI services with Znuny opens up several practical use cases to enhance efficiency and agent productivity. These include ticket classification, where AI can suggest appropriate queues, types, priorities, or dynamic fields based on ticket subject and article text. Information extraction allows AI to convert unstructured customer requests into structured data fields for agents to review. For agent assistance, AI can summarize lengthy conversations or draft initial responses for human approval. Finally, for self-service, a separate AI assistant can be connected to a knowledge base to answer common queries, handing over unresolved requests to Znuny as new tickets. It's recommended to start in suggestion mode and measure accuracy before full automation.

Quellen / Sources

How can I implement automated ticket classification in Znuny using external AI services?

Implementing automated ticket classification in Znuny involves connecting an external AI classifier through a controlled integration layer. The classifier receives only the necessary ticket fields (e.g., subject, article text), processes them, and returns a prediction for fields like queue, type, or priority. The integration layer then validates these predictions against allowed values before writing the results back to Znuny. Before enabling full automation, it's crucial to:

  • Define the specific queues and fields that the AI is permitted to modify.
  • Thoroughly test the classifier using representative historical or synthetic tickets.
  • Establish clear confidence thresholds and a fallback queue for uncertain predictions.
  • Log all predictions, manual overrides, and any processing errors for auditing.
  • Regularly review the quality and accuracy of the classifications as ticket patterns evolve.

OpenTicketAI for Znuny is one documented option for on-premise classification, or a custom service can utilize the Znuny REST API.

Quellen / Sources

How do chatbots and virtual assistants integrate with Znuny, and what are their typical functions?

Chatbots and virtual assistants are typically separate applications that integrate with Znuny, often via the customer portal or the Znuny REST API. They are not natively part of the Znuny core system. Their functions depend heavily on the chosen provider and the specific integration, but generally include:

  • Answering approved FAQs: Providing instant responses to common customer inquiries by drawing from a knowledge base.
  • Collecting structured details: Guiding users through a series of questions to gather necessary information before creating a ticket.
  • Creating tickets: Automatically generating new tickets in Znuny for agents when a query cannot be resolved by the bot or requires human intervention.

It is critical to ensure that any AI-generated answers or actions do not bypass existing access controls, approval rules, or the defined handover processes within Znuny. The exact capabilities and integration methods will vary based on the external service used.

Quellen / Sources

What are the key privacy and operational considerations when using AI with Znuny?

When integrating AI with Znuny, privacy and operational security are paramount due to the sensitive nature of ticket data. Ticket text can contain personal, contractual, or security-sensitive information, making careful handling essential. Key considerations include:

  • Data Minimization: Transmit only the absolutely necessary ticket fields to external AI services to reduce exposure.
  • Data Retention: Document and enforce clear policies for how long AI services retain inference data and logs.
  • Credential Restriction: Strictly limit and manage the credentials provided to AI services, adhering to the principle of least privilege.
  • Data Storage Location: Verify where AI inference and logs are stored, especially for compliance with data residency requirements.
  • On-premise vs. Cloud: For regulated environments, an on-premise AI design can significantly reduce external data transfer, though it still requires robust patching, monitoring, backups, and access controls.

Operationally, it's vital to track metrics like classification precision, manual overrides, routing failures, and the time to first response to continuously assess the AI's impact and reliability.

Quellen / Sources

Are there specific AI add-ons or integrations available for Znuny?

Yes, while Znuny itself doesn't include a native AI suite, several add-ons and integration options exist to bring AI capabilities to your service desk. For instance, OpenTicketAI for Znuny is a documented option for on-premise ticket classification, allowing you to manage AI processing within your own infrastructure. Another notable example is the Znuny-LLM Add-on, which provides self-hosted Large Language Model (LLM) assistance for various tasks, including multilingual summaries, FAQ drafting, and prompt-injection detection, directly within Znuny. Additionally, partners like Sector Nord AG offer an "AI add-on for Znuny" that enables AI-supported automation, powerful chatbots, and intelligent workflows. For custom integrations, the robust Znuny REST API can be used to connect virtually any external AI service or model. These solutions allow organizations to leverage AI for tasks like classification, information extraction, and agent assistance, enhancing Znuny's capabilities without altering its core function as a system of record.

Quellen / Sources