Token Factory Use Cases
From internal knowledge retrieval to embedding AI capabilities into your own products — four anticipated scenarios covering how it works, bloomax's scope of responsibility, and what to confirm before deployment.
The use cases presented on this page are anticipated examples based on the current concept. They do not represent completed deployments or currently available features.
Internal Knowledge Retrieval
Inference infrastructure for an AI assistant that references internal policies, manuals, and other documents to generate draft answers to employee questions. Designed for use cases that support the work of finding the right information.
How it works
A search and retrieval function on your side pulls documents relevant to the question and passes that information to the AI model to generate a draft answer. The configuration combines document search and access control management with the inference processing that generates responses.
Ask a question
Enter what you want to know
Search internal documents
Retrieve relevant content based on access permissions
Generate a draft answer
Generated from the question and search results
Display answer with sources
User reviews the source references
- STEP 1
Ask a question
Enter what you want to know
Your application - STEP 2
Search internal documents
Retrieve relevant content based on access permissions
Your application - STEP 3
Generate a draft answer
Generated from the question and search results
bloomax inference - STEP 4
Display answer with sources
User reviews the source references
Your application
Document search, access control, and source attribution are implemented on the application side. The inference layer generates draft answers from the information it receives.
Things to confirm before deployment
- Types, volume, and update frequency of target documents
- Access permissions by department or user
- Number of concurrent users and required response speed
Customer Support
Inference infrastructure to support inquiry classification, draft response generation, and interaction history summarization. Designed for use cases where staff review AI output as part of their daily support workflow.
How it works
Inquiry content along with relevant FAQs and product information is passed from your application to the inference infrastructure. Staff review and edit the generated draft before use, and interaction history is summarized as needed.
Receive an inquiry
Organize the inquiry and relevant information
Generate a draft response
Generated from FAQs and product information
Review the draft
Staff checks and edits the response
Send the response
Approved content is sent
- STEP 1
Receive an inquiry
Organize the inquiry and relevant information
Your application - STEP 2
Generate a draft response
Generated from FAQs and product information
bloomax inference - STEP 3
Review the draft
Staff checks and edits the response
Staff - STEP 4
Send the response
Approved content is sent
Your application
Staff review AI-generated drafts before use. Response delivery and approval workflows are designed individually.
Things to confirm before deployment
- Types of inquiries handled and languages used
- How FAQs and product information are managed
- Personal data handling and staff review procedures
Large-Scale Document Processing
Inference infrastructure for batch summarization, information extraction, and classification of reports and business documents. Designed for use cases that organize regularly accumulated documents into a format that is easier to work with.
How it works
Documents are prepared in a processable format and passed to the inference infrastructure along with task-specific instructions. Output is reviewed, reprocessed as needed, and then used for ingestion into business systems.
Prepare documents
Convert to a processable format
Summarize, extract, classify
Process documents for the intended purpose
Review results
Check output and reprocess as needed
Use in operations
Feed confirmed results into business systems
- STEP 1
Prepare documents
Convert to a processable format
Your application / integration - STEP 2
Summarize, extract, classify
Process documents for the intended purpose
bloomax inference - STEP 3
Review results
Check output and reprocess as needed
Staff - STEP 4
Use in operations
Feed confirmed results into business systems
Your application
OCR, file conversion, and integration with business systems are confirmed separately from inference processing.
Things to confirm before deployment
- Document format, volume, and length per document
- Target processing time and execution frequency
- Output format, required accuracy, and result review process
AI Integration into Your Own Products
Inference infrastructure for embedding text generation and conversational AI capabilities into SaaS and business applications. Designed for delivering AI features tailored to your own screens and workflows.
How it works
Your application sends requests to the service interface and receives inference results to display in the UI. Configuration and operational requirements are determined based on usage variability and peak load.
Use an AI feature
User inputs or interacts
Send a request
Required information sent to the service
Process with AI model
Generate results from the input
Display results
Reflected in your product's interface
- STEP 1
Use an AI feature
User inputs or interacts
End user - STEP 2
Send a request
Required information sent to the service
Your application - STEP 3
Process with AI model
Generate results from the input
bloomax inference - STEP 4
Display results
Reflected in your product's interface
Your application
Supported models, API specifications, and compatibility with existing systems are confirmed individually.
Things to confirm before deployment
- Model to be used and input/output specifications
- Concurrent request count, peak workload, and response speed
- Credential management and error handling
Notes on scope and individual terms
bloomax's scope of responsibility for each use case is defined by the service arrangement and individual contract. Application development, data integration, and external system connectivity are confirmed separately from the inference infrastructure scope.
Discuss your use case
You can reach out even if your use case or scale is not yet fully defined.