Spaces: Your Knowledge Hubs

Spaces are the foundation of organizing and segmenting information within Spinf. Think of them as dedicated digital rooms, each designed to house specific knowledge and configurations for different aspects of your business.

What are Spaces?

Spaces are versatile containers that allow you to:

• Organize your knowledge base

• Segment information for different products, services, or campaigns

• Apply specific configurations and prompts

Purpose of Spaces

The primary purposes of Spaces are:

1. Focused Knowledge Management: Each Space can be dedicated to a particular product, service, offer, or marketing campaign. This focused approach ensures that the right information is readily available when needed.

2. Tailored Assistance: By associating a Space with a session (e.g., a customer call or presentation), you ensure that the AI assistant has access to the most relevant information and configurations for that specific context.

3. Team Collaboration: Spaces can be shared among team members, fostering collaboration and ensuring consistency in customer interactions across your organization.

4. Efficient Resource Utilization: By segmenting your knowledge base into Spaces, you optimize the use of the AI's processing power, focusing on the most relevant information for each interaction.

Adding Content to Spaces

Spinf offers flexible options for populating your Spaces with diverse content:

1. Extension Pop-up Menu: Easily add content from any web page you're browsing using the Spinf browser extension.

2. Web App Interface: In the Spaces section of the Spinf web application, you can directly upload or link content.

3. Wide Range of Sources: Spinf supports content from various sources, including:

   • Web pages

   • Wikis

   • PDF documents

   • Google Slides presentations

   • Intranet pages

   • CRM systems (e.g., Salesforce)

This versatility ensures that you can incorporate all relevant information into your Spaces, regardless of where it's stored.

Auto RAG and Indexing Features

Spinf employs advanced Retrieval-Augmented Generation (RAG) and indexing techniques to manage large amounts of information efficiently:

1. Handling Large Datasets: The system can load and process more information than a typical LLM (Large Language Model) can handle in a single call.

2. Smart Content Processing:

   • Spaces below 25K tokens are used as-is by the LLM.

   • Larger spaces automatically utilize RAG and indexing for efficient information retrieval.

3. Token Measurement: The interface provides token counts for each Space, displayed as both absolute values and as a proportion of the total context size.

Space Management Recommendations

While there's no strict limit to the size of a Space, we recommend the following best practices:

1. Avoid Overloading: Rather than cramming too much information into a single Space, it's better to segment information across multiple Spaces.

2. Focused Content: Keep each Space focused on a specific topic, product, or service for optimal performance.

3. Regular Review: Periodically review and update your Spaces to ensure the information remains relevant and current.

4. Balancing Act: Aim for a balance between comprehensive information and manageable size. Spaces below 25K tokens are used directly by the LLM, which can lead to faster processing in some cases.

By following these guidelines and leveraging Spinf's powerful Space management features, you can create an optimized knowledge structure that enhances the AI's performance and provides more accurate, contextually relevant assistance during your customer interactions.