Skip to main content
This page explains the fundamental concepts behind RevBridge AI. Understanding these will help you get the most out of the platform.

Multi-Armed Bandit (MAB) Campaigns

At the heart of RevBridge AI is the Multi-Armed Bandit algorithm—a machine learning approach that balances exploration and exploitation to optimize campaign performance.

How It Works

Traditional A/B testing requires you to:
  1. Run a test for a fixed period
  2. Wait for statistical significance
  3. Manually pick a winner
  4. Apply the winning variant
MAB does this continuously and automatically:
  1. Exploration: Tests different variations (templates, channels, send times)
  2. Exploitation: Gradually shifts traffic to better-performing variations
  3. Adaptation: Continuously adjusts as performance data changes
The result? Your campaigns automatically optimize themselves, finding the best combination of message, channel, and timing for each audience segment.

What Gets Optimized

Creative Concept Framework

RevBridge AI uses a structured framework of 6 Creative Concepts to generate diverse template variations: When you generate templates, the AI creates variations using each concept, then the MAB algorithm determines which resonates best with your audience.

Customer Data Platform (CDP)

RevBridge AI includes a built-in Customer Data Platform that unifies customer data from multiple sources.

Customer Profile

Each customer in RevBridge AI has a unified profile containing:
  • Identity: Email, phone, name, and other identifiers
  • Attributes: Custom properties like preferences, segments, tags
  • Events: Behavioral data like page views, purchases, cart additions
  • Purchases: Complete transaction history with items and revenue
  • Engagement: Email opens, clicks, SMS responses, etc.

Data Sources

Customer data flows into RevBridge from:
  • Integrations: Shopify, VTEX (coming soon)
  • Events API: Custom events from your application backend
  • Web SDK: Page views, sessions, and events captured in the browser
  • CSV Imports: Manual data uploads
  • Campaign Interactions: Engagement data from RevBridge campaigns

Audiences (Segments)

Audiences are groups of customers that share common characteristics. They’re used to target campaigns and analyze behavior.

Audience Types

Dynamic vs. Static

  • Dynamic Audiences: Membership updates automatically as customer data changes
  • Static Audiences: Fixed list of customers (e.g., CSV imports)

Brand Dossier

A Brand Dossier is RevBridge AI’s understanding of your brand identity. It includes:
  • Visual Identity: Colors, typography, logo usage
  • Brand Voice: Tone, personality, communication style
  • Messaging Guidelines: Preferred terms, phrases to avoid
  • Accessibility: WCAG compliance requirements

Brand Intelligence

When you scan your website, RevBridge AI uses Brand Intelligence to automatically extract:
  1. Color Palette: Primary, secondary, accent, and neutral colors
  2. Typography: Font families, sizes, and hierarchy
  3. Visual Style: Photography style, imagery preferences
  4. Logo: Primary logo for use in templates
This information guides the AI when generating templates, ensuring consistency with your brand.

Templates

Templates are the message content sent to customers. RevBridge AI generates templates across multiple channels:

Channel Formats

Template Variables

Templates use variables (placeholders) that get replaced with customer data:
Common variables:
  • {{first_name}} - Customer’s first name
  • {{headline}} - Dynamic headline (set by campaign)
  • {{body}} - Main message content
  • {{cta_text}} - Call-to-action button text
  • {{cta_url}} - Call-to-action destination URL

Campaigns

A Campaign in RevBridge AI is an automated marketing initiative that sends messages to an audience.

Campaign Types

Campaign States

Campaign Metrics

Key metrics tracked for each campaign:
  • Reach: Total customers targeted
  • Delivered: Messages successfully delivered
  • Opens: Email/push opens (where trackable)
  • Clicks: Link clicks
  • Conversions: Completed desired actions
  • Revenue: Attributed revenue from conversions
  • CPA: Cost per acquisition
  • Lift: Improvement vs. control group

Events

Events are actions or occurrences tracked for customers. They power behavioral segmentation and campaign triggers.

Event Structure

Each event contains:
  • Event Type: Category of event (e.g., purchase, page_view)
  • Timestamp: When the event occurred
  • Properties: Additional data about the event
  • Customer ID: Which customer performed the action

Common Event Types

Integrations

Integrations connect RevBridge AI to external platforms for data synchronization.

Integration Flow

  1. Authentication: OAuth-based secure connection
  2. Initial Sync: Historical data import
  3. Incremental Sync: Ongoing data updates
  4. Bi-directional (where supported): Push data back to source

Supported Integrations

Channels

Channels are the communication methods used to reach customers.

Channel Capabilities

Push, RCS, and Voice channels are coming soon. Email, SMS, and WhatsApp are available now.

Channel Selection

The MAB algorithm automatically determines the optimal channel for each customer based on:
  • Historical engagement on each channel
  • Message urgency and content type
  • Customer preferences and availability
  • Channel deliverability and cost