'/%3e%3cpath%20id='cross_2'%20data-name='cross%202'%20class='cls-2'%20d='M20.612,30.43h0a5.361,5.361,0,1,0-10.723,0H0A33.709,33.709,0,0,1,30.569,0V9.75a5.361,5.361,0,1,1,0,10.723V30.43Z'%20transform='translate(67.193%200.007)%20rotate(90)'/%3e%3cpath%20id='cross_3'%20data-name='cross%203'%20class='cls-3'%20d='M20.612,30.43h0a5.361,5.361,0,1,0-10.723,0H0A33.709,33.709,0,0,1,30.569,0V9.75a5.361,5.361,0,1,1,0,10.723V30.43Z'%20transform='translate(67.186%2067.041)%20rotate(180)'/%3e%3cpath%20id='cross_4'%20data-name='cross%204'%20class='cls-4'%20d='M20.6,30.5h0a5.361,5.361,0,1,0-10.723,0H0A33.71,33.71,0,0,1,30.56,0V9.818a5.361,5.361,0,1,1,0,10.723V30.5Z'%20transform='translate(0%2067.041)%20rotate(-90)'/%3e%3c/svg%3e)
AI Features FAQ
When considering the adoption of generative AI, data security is usually the first concern that comes up. SetFN offers AI features in two ways: local AI and external AI. Local AI keeps data from being transmitted outside the organization, making it suitable for security-sensitive environments. This document answers frequently asked questions related to AI features.
Local AI vs External AI
Q: What is the difference between local AI and external AI?
Local AI runs models with an AI runtime (such as Ollama or LM Studio) on the SetFN server or a server on the same network. External AI works by calling the API of an external service such as OpenAI or Anthropic.
| Item | Local AI | External AI |
|---|---|---|
| Data transmission | Does not leave the organization's equipment | Sent to an external service |
| Internet required | Not required | Required |
| Response quality | Depends on the model | Can use large-scale models |
| Cost | Server operating cost only | Billed per API call |
| Security | Managed by your organization | Depends on the external service's policies |
Q: Which approach should I choose in a security-sensitive environment?
You should use local AI. With local AI, all processing happens on your organization's equipment, so data is never transmitted outside the organization. If you prepare the model files in advance, it can even be used in closed network environments without an internet connection.
Q: Can local AI and external AI be used at the same time?
Yes. You can configure both local AI and external AI and choose between them depending on the situation. For example, you could use local AI for security-sensitive documents and external AI when you need higher-quality responses.
Data Security
Q: If I use AI features, is my data used for training?
Because local AI runs on your organization's equipment, data never leaves the organization and is not used for model training. Only when you run training yourself with the fine-tuning plugin is the selected data used for training. If you use external AI, data handling follows the policies of that external service.
Q: What scope of data is sent to the AI?
The AI receives the user's question along with the file content or search results needed to handle the request. Context search and AI work requests use related documents found in the knowledge index as the basis for answers. If you use external AI, this content is sent to the external AI provider. External AI connected through MCP accesses documents, search terms, paths, and activity history within the scope the administrator allows and according to the user's permissions.
Hardware Requirements
Q: Is a GPU required to use local AI?
You can use local AI with CPU only, without a GPU. However, having a GPU significantly speeds up response times.
| Environment | Response speed | Recommendation |
|---|---|---|
| CPU only | Slow (tens of seconds or more) | Suitable for light tasks |
| GPU equipped (8GB+ VRAM) | Fast (within a few seconds) | Recommended for general work |
| High-performance GPU (16GB+ VRAM) | Very fast | Recommended for large document processing |
Q: What are the minimum server specs needed for local AI?
This varies depending on model size, but in general the following specs or higher are recommended.
- RAM: 16GB or more
- Storage: 10GB or more of free space for AI model files
- GPU (optional): NVIDIA GPU, 8GB+ VRAM
Supported Models and Features
Q: What AI models can be used?
For local AI, you can use models that run on local AI runtimes such as Ollama, LM Studio, and Docker Model Runner. For external AI, you can connect APIs such as OpenAI, Anthropic, or Gemini.
Q: What tasks can AI perform?
SetFN's AI features let you do the following.
- AI work requests: Ask in chat to search, add schedule events, summarize work, or draft emails
- Context search: Find files with similar meaning even without an exact keyword
- Image search: Search for text inside images or for similar images
- Editor AI autocomplete: Get suggestions for what comes next while writing a text document
- Mail AI assistant: Summarize mail and organize to-dos and phishing indicators
- Automatic analysis: When files are created or modified, local AI generates metadata such as summaries and tags
For details on each feature, see AI Features Summary.
Q: What file formats does the AI feature support?
The AI can read and analyze the content of text-based files (TXT, MD, CSV, JSON, etc.) and document files (PDF, DOCX, etc.). Images can be read and analyzed with a vision model (VL) or OCR. Other binary files such as video are not included as AI analysis targets.