Support tickets
Classify queries by topic, intent and priority. Define labels linked to your queues and how to handle tickets with multiple issues.
Use cases / Text and agents
Text classification
Classify text with AI and organize incoming requests.
Turn tickets, emails and documents into categories your application can use. Detect intent, assign labels and direct each request to the right team or process.
With QDivZero, you can run an LLM or a specialized text classification model. Define your categories, connect the model through an API and automate content organization with your business rules.
From text to a decision
Organize large volumes of content automatically.
Turn messages and documents into useful categories. Connect labels to your application’s queues and workflows.
Documents and files
Organise files by content. Provide review categories for incomplete documents or those outside the expected types.
Emails and messages
Identify sales enquiries, incidents and reasons for contact. Evaluate short messages, abbreviations and languages found in your inbox.
Request routing
Use labels to prepare assignment to teams and workflows. Your application applies the rules and handles ambiguous cases.
Build it with QDivZero
From a text input to the next process step.
Define labels and examples, then send each text to a compatible model. Your application validates categories and applies assignment rules, including a route for ambiguous cases.
Your application flow
Text
Receive the ticket, email, document, or message.
Categories
Define labels, criteria, and classification examples.
Model
The classifier or LLM assigns the category.
Process
Your application validates the label and selects the next step.
Open-weight / Hugging Face
You can start with…
An LLM lets you describe categories and provide examples; a specialized model can fit a stable, repetitive classification task. Compare accuracy by category, language support and volume to choose your deployment. Use the same categories and a separate evaluation set to compare an LLM with a specialist model. Check less frequent classes and the cost of serving expected volume.
Text and vision
Qwen3.8-27B
For conversation, code, and tasks combining text, images, and your own context.
View model on Hugging FaceText and vision
GLM-5.3 Flash
Combine conversation, visual understanding, and work with tools.
View model on Hugging FaceYour own models
Your own fine-tuned model
For a defined task with your own categories, examples, and training data.
Deploy your modelText classification
Text classification: frequently asked questions
What is AI text classification?
Text classification assigns labels or categories to messages, emails, and documents based on content. Identify topics, intents, or priorities to organize information and automate process decisions. Use an LLM or a specialist model.
How do I automatically classify support tickets?
Define categories and example tickets for each. Send text to the model and request a label in the format your system needs. Your application validates the output and assigns the request to the appropriate team or process.
Can AI detect intent in emails and messages?
Yes. Define intents such as sales inquiry, issue, or information request and ask the model to identify the category. Add business examples and a category for cases requiring review or outside your defined labels.
Should I use an LLM or a fine-tuned model for text classification?
An LLM lets you start with instructions and adjust categories quickly. A fine-tuned classifier may suit recurring tasks with stable labels and your own examples. Compare both using representative data, category quality, and required capacity.
Can open-weight models classify text in different languages?
Yes, with a model supporting your language and performing well on your categories. Start by comparing Qwen3.8 or GLM-5.3 Flash on real examples. Include domain expressions, short messages, and ambiguous cases when choosing a configuration.
How do I integrate text classification through an API?
Deploy a compatible model on QDivZero and send text from your application. Receive a category in a defined format and connect it to business rules. Organize individual requests or batches based on volume and deployment capacity.
Can AI analyze sentiment in text?
Yes. Define sentiment categories and ask an LLM to classify messages, or use a model trained for the task. Evaluate results with your business languages and text, especially when messages include irony, ambiguity or specialized vocabulary.
How do I improve automatic classification accuracy?
Define distinct categories and provide examples for each. Evaluate real messages, review confusion between classes and allow an output for ambiguous cases. If you need multiple labels, specify that format and validate results before using them to route requests.
Ready to organize your messages with AI?
Automate ticket, email and document classification using your own categories. Deploy a compatible model in QDivZero and connect its labels with your application’s queues, teams and processes.