Admin Console Overview
Welcome to Control Bridge! This guide will introduce you to the admin interface and help you understand how to manage your AI-powered email processing system.
Overview
Control Bridge is a comprehensive web-based administration interface for managing your email automation system. It provides full control over:
- AI Configuration - LLM providers, models, specialist agents, and tools
- Email Processing - Monitored mailboxes, subscriptions, and VIP contacts
- Organization Data - Companies, departments, employees, and teams
- Search Integration - Knowledge base providers and indexes
- Activity Monitoring - Execution logs, metrics, and debugging
- Safety Gateway - Outbound communication security with PII detection and content analysis
The console is designed for system administrators who need to configure and monitor the email processing system.
Your 30 Day Trial
New organizations start on the 30 Day Trial tier. During the trial you have access to everything the Pro Plus plan offers, including the full agent lineup, integrations, and organizational knowledge features, plus $100 in LLM credits (currently available to US customers) so you can try things out without connecting your own AI provider. The trial requires no credit card; when it ends, choose a paid plan under Administration > Settings > Subscription to keep going. Feature docs that say "Pro Plus only" apply to paid plans; those features are included while you are on the trial.
System Architecture
Outermind uses a sophisticated multi-agent architecture to process incoming emails intelligently:
Email Received
↓
Microsoft 365 Shared Mailbox
↓
Graph API Webhook → Outermind API
↓
SuperAgent (Routes to appropriate specialist)
↓
Specialist Agent (Processes with assigned tools)
↓
LLM Processing (Anthropic, OpenAI, or Azure)
↓
Tool Execution (Reply, Search, Escalate, etc.)
↓
Response Sent / Action Taken
The SuperAgent System
Outermind uses a two-tier agent architecture:
- SuperAgent - Routes incoming emails to the most appropriate specialist agent based on email content, sender, and context
- Specialist Agents - Execute domain-specific processing using assigned tools and LLM reasoning
This architecture allows you to create specialized agents for different purposes:
- Customer support
- Scheduling and calendar management
- HR inquiries
- Sales leads
- Technical support
- And more...
Each specialist agent:
- Has custom instructions (system prompt) that define its behavior
- Uses a specific LLM model (e.g., Claude Sonnet, GPT-4)
- Has access to assigned tools (reply, search, escalate, etc.)
- Tracks costs and performance metrics
Main Hubs
The Admin Console is organized into four operational hubs accessible via tabs in the header:
- Workforce (
/workforce) - The AI agents themselves: agent fleet, personal assistants, monitored mailboxes, AI providers, and agent self-service - Capabilities (
/capabilities) - What the workforce can do: skills, knowledge, tools, integration connections, and development sandboxes - Governance (
/governance) - Oversight of the workforce: live operations, activity history, auditing, safety, and data governance - Administration (
/administration) - Account administration: key contacts, activity summary, and account settings
Two executive hubs appear alongside them when available: Boardroom (/boardroom) and Personal Assistant (/pa-hub). Account Settings (/account-settings) and the Support Center (/support) are reached from the profile avatar menu in the top-right. The home page (/) is a separate landing page you can return to any time by clicking the Outermind logo.
Workforce
The Workforce hub is home to the AI agents themselves and the intelligence behind them.
Agents
- Agents & SMEs - Create and manage specialist agents and Subject Matter Experts with custom instructions
- Personal Assistants - Manage the per-employee AI assistants
- Monitored Mailboxes - Add mailboxes to monitor, assign agents, configure auto-reply
- Agent Index & Search - Browse and search the agent capability index
- DreamState - Configure background self-improvement runs
Intelligence
- AI Providers - Add and configure AI services (Anthropic, OpenAI, Azure OpenAI, Azure AI Foundry) and their models
- Execution Profiles - Tune model tiers and execution settings
Agent Self-Service
- Creation Settings - Configure self-service agent creation
- Approvals - Review agent-creation approval requests
- Templates - Manage agent-creation templates
When to use: Create agents, connect AI providers, manage mailboxes, configure agent self-service
Capabilities
The Capabilities hub covers what the workforce can do.
Skills
- Skills - Create and manage reusable agent skills
Knowledge
- Knowledge - Configure knowledge sources, knowledge bases, indexes, and scans (tabbed page)
Research
- Company Research - Configure automated company research
Tool Management
- All Tools - Browse every tool available to agents, built-in tools included
Connections
- + Add Connection - The integration marketplace; connected integrations (HubSpot, GitHub, LinkedIn, QuickBooks, SQL Tools, API Tools, Team Messaging, and more) appear here once configured
Development Sandboxes
- On-demand container sandboxes for code execution (shown when your subscription includes them)
When to use: Configure tools, connect integrations, manage knowledge and skills
Governance
The Governance hub provides oversight of the AI workforce.
Live Operations
- Pending Actions - Combined review queue for approvals and safety reviews
- Escalations - Handle items escalated for human attention
- System Diagram - Interactive architecture diagram with real-time updates via SignalR
Activity History
- Agent Executions - View execution history with metrics, costs, and status
- Agent-2-Agent Comms - Trace inter-agent messages
- SME Activity - Track subject matter expert requests and invocations
- Platform Usage - Usage and cost analytics
- Execution Stats - Aggregate execution statistics
Auditing
- Integration Logs - View third-party integration activity logs
- Agent Creation Logs - Audit self-service agent creation
Configuration
- Safety Gateway - Configure risk thresholds, gateway modes, and LLM analysis options, and review the safety audit log
- Escalation Router - Configure escalation routing
- Retrospectives - Configure agent retrospectives
- PA Policies - Personal Assistant privacy and policy settings
- Meeting Intelligence - Configure meeting transcription and follow-up
- Microsoft Entra Agents - Manage Agent 365 identities
AIPSO (shown when entitled)
- The AI Platform & Security Officer: Overview, Settings, Findings, Approvals, Config Log, and Chat
Data & Logs
- Storage - Bring-your-own-storage settings
- Memories - View and manage Personal and Global Memories
- Agent Artifacts - View agent-generated artifacts and cached content
- Scheduled Reminders - Manage scheduled reminders
When to use: Monitor system health, review queues, audit activity, configure safety and governance
Administration
The Administration hub handles account administration.
Account
- Key Contacts - Manage organization key contacts
- Activity Summary - Review the tenant activity summary
- Settings - Configure system-wide settings, pause controls, and agent partnerships
When to use: Manage account settings, key contacts, and pause controls
Account Settings and Support Center
Two more areas are reached from the profile avatar menu in the top-right, not from the hub tabs:
- Account Settings (
/account-settings) - Subscription and billing, Data Privacy, team Users and Roles, and organization data (Companies, Key Contacts, Company Research, Provisioning API Keys) - Support Center (
/support) - Support dashboard, your tickets, new ticket submission, and the notifications inbox
Permissions & Access
Authentication
The Admin Console requires Azure AD authentication. You must be signed in with a Microsoft 365 account that has admin access.
Required Permission
All admin console features require the system.admin permission. This ensures only authorized administrators can configure the system.
Security
- All API calls are authenticated via Azure AD tokens
- Sensitive credentials (API keys) are stored securely in Azure
- Activity is logged for audit purposes
User Interface Patterns
Responsive Design
The console adapts to your screen size:
- Desktop - Full three-column layouts with sidebar, content, and details
- Tablet - Two-column layouts with collapsible sidebar
- Mobile - Single column with drawer navigation
View Modes
Many pages offer table view and card view:
- Table View - Compact, sortable columns (best for desktop)
- Card View - Visual cards with key information (best for mobile)
Use the toggle button in the top-right of data pages to switch views.
Form Pages and Modals
Create and edit operations use a mix of approaches:
- Form Pages - Complex configurations like Agents, SMEs, Search Providers, and Search Indexes use dedicated form pages for a better editing experience
- Modal Dialogs - Simpler operations use modal dialogs that overlay the current page for quick actions
Common Actions
- Create - Blue "+" button or "Add New" button
- Edit - Pencil icon or click the row/card
- Delete - Trash icon (usually with confirmation)
- Search/Filter - Search box and filter controls at top of lists
- Pagination - Controls at bottom of tables (rows per page, page navigation)
Key Features
AI-Powered Tools
The console includes AI-powered features to help you:
- Prompt Improvement - AI suggests better agent instructions when creating/editing agents
- Execution Troubleshooting - AI analyzes failed executions and suggests fixes
- Auto-Thread Detection - Automatically includes email thread context for recent conversations
Real-Time Updates
The dashboard system diagram updates in real-time using SignalR websockets. Configuration changes appear immediately without page refresh.
Cost Tracking
Every agent execution tracks:
- Token usage (input and output)
- LLM costs (based on configured model pricing)
- Execution duration
- Tool call success/failure
View costs per execution or aggregate by agent, model, or time period.
Azure AD Integration
Seamlessly integrates with Microsoft 365:
- Azure AD authentication
- Employee directory sync
- Team/group management
- Shared mailbox access via Microsoft Graph API
What's Next
Now that you understand the Admin Console at a high level, here are recommended next steps:
- Learn Navigation - Master the console interface
- Quick Start Guide - Complete your first workflow
- Create Agents - Set up specialist agents
- AI Chief of Staff - Deploy an autonomous AI executive for strategic goal management
- Personal Assistant - Give every employee a dedicated AI assistant for email, goals, and productivity
- Monitor Activity - Review execution logs
- Configure Safety Gateway - Protect outbound communications
- Agent Intelligence - Enable memories and artifacts for smarter agents
Getting Help
If you encounter issues or have questions:
- Support Center - Submit tickets and track support requests at
/support - Troubleshooting Guide - See Common Issues
- Debugging Techniques - See Debugging Guide
- Documentation - Browse all articles using the sidebar navigation
Summary
Control Bridge provides comprehensive control over your AI-powered email processing system. Organized into four operational hubs—Workforce, Capabilities, Governance, and Administration, with Account Settings and the Support Center in the profile menu—you have all the tools needed to manage and optimize your automated email workflows.
The SuperAgent architecture enables sophisticated email routing and processing, while detailed execution logs and cost tracking give you full visibility into system performance. The Safety Gateway adds enterprise-grade security by analyzing all outbound communications before they are sent.
Ready to get started? Continue to the Navigation Guide to learn how to use the interface effectively.