Discussion and asking suggestion about my idea A developer is soliciting architectural feedback for a proposed multi-tenant platform that would let businesses deploy AI-powered customer support using their own documents, policies, and product data. The open questions center on whether to isolate vector databases per tenant or share a multi-tenant vector store with metadata-based separation, how to organize frequently changing knowledge, and whether to deliver the product as an embeddable chatbot, SDK, API, or a combination. πŸš€ Exploring an idea β€” would love to hear your thoughts I’ve been thinking about building a platform that can help businesses provide AI-powered customer support using their own business data, documents, policies, product information, and more. But one question I’m currently exploring is not the AI model itself β€” it’s the architecture and delivery. For a platform that may serve multiple businesses: πŸ”Ή Should we maintain completely separate vector databases for each business? πŸ”Ή Or would a shared, multi-tenant vector architecture with proper metadata isolation be more efficient and scalable? πŸ”Ή How should the knowledge be organized when a business has products, policies, FAQs, documents, and constantly changing information? πŸ”Ή More importantly, how should the final product actually be delivered to the client? Should it be: β†’ A simple website-embeddable chatbot β†’ An SDK/component they can integrate into their existing application β†’ An API-first platform β†’ Or a combination of these approaches? I’m currently exploring the best balance between scalability, data isolation, infrastructure cost, ease of integration, and the actual value delivered to businesses. I’d genuinely appreciate perspectives from developers, architects, SaaS builders, and anyone who has worked with RAG, vector databases, multi-tenant systems, or AI-powered customer support platforms. I’m still shaping the architecture, so I’d love to learn from different approaches and experiences. πŸ™Œ If you have any ideas, suggestions, or architectural approaches, please feel free to comment or DM me. I’d be really excited to discuss and learn from you πŸš€