PLai Architecture
- Organization Level
- Organizations (Teams)
- Projects (Workspaces)
- Members & Roles
- Organization Settings
- Invitations
- Billing
- Team Management
- Project Level (Inside Each Project)
- Agents (AI Workers)
- Tools (Service Integrations)
- Datasources (Knowledge Bases)
- Workflows (Multi-step Automation)
- Jobs (Scheduled Execution)
- Triggers (Event-based Automation)
- Monitor (Project Analytics & Metrics)
- User Level (Global)
- Profile & Account Settings
The Five Core Concept Groups
Organizations & Projects
Structure, team management, and workspace isolation
Agents, Datasources & Tools
The building blocks you create and configure
Workflows, Jobs & Triggers
How to automate and schedule your work
Security & Compliance
Access control, credentials, and safety guardrails
Monitoring
Project-level analytics, metrics, and performance tracking
Understanding the Hierarchy
Level 1: Organizations & Projects
Scope: Organization-wide structure Where you live:- Organizations - Top-level container for your team
- Projects - Isolated workspaces where you do the actual work
- Members - People with roles (OWNER, ADMIN, VIEWER)
Level 2: Building Blocks
Scope: Inside a Project What you create:-
Agents - AI assistants that perform tasks
- Configuration: Model, system instructions, capabilities
- Properties: Name, description, role-based settings
-
Tools - External services agents can use
- Types: API Requests, Web Search, Browser, MCP Servers, Datasources
- Purpose: Extend agent capabilities
-
Datasources - Knowledge bases for agents
- Types: Files, URLs, Websites, Content
- Purpose: Provide context and facts to agents
Level 3: Automation
Scope: Inside a Project How you scale:-
Workflows - Multi-step sequences combining agents and tools
- Example: Route customer → Sales Agent → CRM Tool → Response
-
Jobs - Scheduled executions
- Example: Run data processing every day at 2 AM
-
Triggers - Event-based automation
- Types: Webhook, Schedule, Manual
- Example: Process form submission automatically
Level 4: Safety & Access
Scope: Spans Organization and Project How you protect:-
Roles & Permissions - Who can do what (RBAC)
- Organization roles control team access
- Project roles control resource access
-
Credentials - Securely store secrets
- API keys, passwords, tokens
- Encrypted storage, audit logged
-
Guardrails & Answer Filters - Safety constraints
- Content filtering
- Rate limiting
- Compliance rules
Level 5: Observability
Scope: Project-level How you understand what’s happening:-
Metrics - Performance data
- Speed, accuracy, success rates, errors
-
Analytics - Usage patterns
- Agent activity, tool usage, workflow performance
-
Activity Logs - Audit trail
- Who did what, when, and what changed
Decision Tree: Which Concept Do I Need?
I want to organize my team ↓ → Organizations & Projects I want to create AI agents ↓ → Agents, Datasources & Tools I want to automate workflows ↓ → Workflows, Jobs & Triggers I want to control access or store secrets ↓ → Security & Compliance I want to track performance and costs ↓ → MonitoringQuick Reference: Typical User Journeys
Journey 1: Getting Started
- Understand structure → Organizations & Projects
- Invite your team → Organizations & Projects (Roles section)
- Create first agent → Agents, Datasources & Tools
- Add knowledge → Agents, Datasources & Tools (Datasources section)
Journey 2: Going to Production
- Set up security → Security & Compliance
- Configure guardrails → Security & Compliance (Guardrails section)
- Monitor performance → Monitoring
- Audit access → Monitoring (Logs section)
Journey 3: Scaling Up
- Build workflows → Workflows, Jobs & Triggers
- Schedule jobs → Workflows, Jobs & Triggers (Jobs section)
- Set up triggers → Workflows, Jobs & Triggers (Triggers section)
- Track results → Monitoring
Key Mental Models
Model 1: Container Nesting
- Organization
- Project 1
- Agent A
- Tool X
- Datasource D
- Project 2
- Agent B
- Agent C
- Tool Y
- Project 1
Model 2: Separation of Concerns
- Org Level (Who + What Access)
- Organizations & Projects
- Members & Roles
- Organization Settings
- Project Level (What You Build & Run)
- Agents, Tools, Datasources
- Workflows, Jobs, Triggers
- Monitor (Performance)
- Security (Spans Both)
- Credentials
- Guardrails
- Activity Logs
Model 3: Data Flow
- Datasource (Knowledge) and User Input feed into the Agent
- Agent → Tool → External Service — filtered/validated by Guardrails along the way
- Result → Monitor (Track) → Analytics Dashboard
Important Reminders
Monitor from day one. It’s easier to understand costs and performance from the start than to debug problems later.
Next Steps
- Understand your structure → Read Organizations & Projects
- Create your first agent → Go to Agents, Datasources & Tools
- Follow the tutorial → Check out Create Your First Agent
- Organizations & Projects - Team structure, roles, workspace management
- Agents, Datasources & Tools - Building your AI workers
- Workflows, Jobs & Triggers - Automating at scale
- Security & Compliance - Protecting your data and access
- Monitoring - Tracking performance and costs