​
Multi-Tenancy

Multi-tenancy allows you to create deeply personalized AI experiences where different users get different agents, knowledge contexts, and combinations of these – all while keeping data completely separate and secure. Think of it as giving each user their own custom AI assistant that knows exactly what information they need access to and responds accordingly.

​
Key Features

  • Each Agent has a stable ID: agentUid
  • Each Knowledge Context has its own ID: contextUid
  • Each Session binds the two: agentSessionUid

​
Programmatic Agent Assignment

  • Assign the same AI agent or different AI agents (HR Agent, Leadership Agent, Technical Agent) to different users
  • Users only see agents relevant to their role and permissions
  • Pre-configure agent behavior, tone, and capabilities
  • Mix and match any agent with any knowledge base programmatically

​
Private Knowledge Contexts

  • Each tenant gets their own private knowledge context
  • Granular control – users only see knowledge relevant to their role
  • Tenants can upload documents, PDFs, and content into your platform
  • Your team creates knowledge contexts based on tenant uploads
  • Real-time content processing and indexing for your users

​
Flexible Tool Integration

  • Different users get access to different tools based on their role
  • Integrate seamlessly with CRM, HR systems, and monitoring tools
  • Build custom workflow tools for specific tenant needs

​
Session-Based Security

  • Every user interaction creates a secure, temporary agent session
  • Sessions bind together the right agent, knowledge, and the MCP server, which contains tools for that specific user

​
Agent Memory & Conversation Continuity

  • Users can return to previous conversations and continue tasks
  • Conversation history is accessible, but new threads start fresh (no cross-thread personalization)

​
Real-Life Example: The Journey in Two Tracks

​
Track A: Manage Knowledge (for Admins)

  • Description
    This track shows how to set up the knowledge management foundation that powers personalized AI experiences. It demonstrates the admin workflow that users need to build to create private, tenant-specific knowledge containers and populate them with relevant content.
  • Summary
    Admins get a simple interface to upload and manage documents that become the AI’s knowledge base. Each tenant gets their own private knowledge context where all uploaded content is processed, indexed, and kept completely isolated from other tenants. This creates the foundation for personalized AI responses based on each organization’s specific information.

​
Steps

  • Design “Manage Agent Knowledge” Screen
    Your product designer creates a simple UI: a screen with an upload button and a list of files. This will let each tenant admin (or end user) upload documents/ content that fuels the agent.
  • Create a Knowledge Context When a new tenant signs up, your developer code calls POST /contexts. This creates a private container (context) in the Mindset AI AMS where that tenant’s content lives. It returns a contextUid.
  • Upload Content Your customer admin uploads HR docs, PDFs, Notion exports, etc. Your developer sends these to PUT /contexts/{contextUid}/files. Mindset AI processes and indexes the content. Once the context is ready, everything in it is private and scoped only to that tenant.

​
Track B: Enable agents to be accessed by different users

Description: This track shows how to connect users with AI agents that have access to their organization’s private knowledge. It demonstrates how to create secure chat sessions that bind together the right agent, knowledge context, and user permissions. Summary: Users get instant access to AI agents that are pre-configured by your product team with the right personality and capabilities, while only being able to access their organization’s private knowledge. Each chat session is secure and temporary, ensuring users only see information they’re authorized to access while maintaining a seamless conversational experience.

​
Steps

  • Configure one shared agent
    Your product manager configures a reusable global agent (tone, steps, actions) in the Mindset AI Agent Management Studio. It gets an agentUid.
  • Start a user session
    When your end user opens the in-app chat, Dev’s code calls POST /agent-sessions with:
    • the shared agentUid
    • the tenant’s contextUid
    • the user’s ID
    This returns a short-lived agentSessionUid, which the front end uses to initialise the chat.
  • Let your end user talk to their private AI
    Your end user chats naturally, and the agent only pulls answers from their company’s indexed docs - thanks to the assigned context.

​
How your end user experiences it (end‑user view)

  1. Clicks Help – the chat window opens.
  2. Asks a question – “How many annual leave days do I have?”
  3. Sees an answer – pulled from the HR policy your customer admin uploaded

​
Use Cases for Your End Users

  1. Role-Based AI Access: Sales team gets CRM-connected agents, while HR gets policy-focused agents with employee data access
  2. Department-Specific Knowledge: Engineering sees technical docs and code repos, while executives see financial reports and strategic plans
  3. Secure Multi-Team Deployment: One AI platform serves your entire organization and different tenants while maintaining complete data separation
Multi-tenancy transforms generic AI into personalized assistants that understand context, remember preferences, and adapt to each user’s unique workflow - all while maintaining enterprise security and easy management. This allows teams to actually provide value to their customers rather than offer a traditional GPT-like experience.