Propensity AI Essentials: Core Concepts – Propensity

What this is

Propensity AI is a conversational interface for creating, managing, and analyzing campaigns within Propensity.

Using natural language prompts, you can create audiences, build campaign strategies, generate messaging, launch campaigns, review performance data, and update existing campaign configurations without manually navigating multiple areas of the platform.

Unlike general-purpose AI tools, Propensity AI can access information already available within your Propensity workspace, including campaign data, audiences, CRM integrations, account and contact intelligence, engagement signals, and workspace settings. This allows the platform to provide recommendations, generate proposals, and assist with campaign management tasks using your existing Propensity data.

Propensity AI also supports custom AI agents. These agents can be configured around specific business requirements, campaign frameworks, audience definitions, brand guidelines, or operational workflows to help teams execute tasks more consistently.

You can access the AI by going to the Propensity menu and clicking Propensity AI.

When to use this

Use Propensity AI when you want to accelerate campaign execution, automate repetitive tasks, or manage campaigns using natural language prompts.

Common use cases include:

Creating campaigns, audiences, and plays from a single prompt
Building audiences using account attributes, contact criteria, intent signals, engagement activity, or CRM data
Generating campaign messaging, ad creative, and email content
Launching campaigns across multiple channels
Creating personalized campaigns for specific accounts, industries, products, regions, or buying circles
Building reusable AI agents for common workflows and campaign processes
Analyzing campaign performance and identifying optimization opportunities
Updating campaign settings, budgets, targeting, messaging, or flight dates
Retrieving campaign, audience, engagement, and pipeline insights without manually navigating reports

Example prompts:

Create personalized campaigns for the top 25 accounts showing purchase intent
Build an EMEA campaign agent that follows regional messaging requirements
Launch separate campaigns for enterprise, mid-market, and SMB audiences
Generate account-specific campaigns using different messaging strategies
Identify engaged contacts and create follow-up campaigns based on their activity

How it Works: The Four Intelligence Layers

Propensity AI translates your natural language inputs into structured execution using four connected operational layers:

Signal Intelligence: Connects your CRM data, web traffic, email interactions, and sales pipeline to surface in-market accounts before external competitors detect them
AI Campaign Execution: Processes a single prompt to build your target group automatically, select the channel mix, write creative copy, and prepare the campaign infrastructure
Omnichannel Reach: Deploys targeting parameters strictly at the contact level (not just account level) across Propensity Display Network (PDN) and external platforms, such as LinkedIn and Meta
Optimization & Attribution: Continuously monitors performance and maps every campaign dollar spent directly back to pipeline revenue generation

Core Capabilities and Actions

Users can interact with the platform to perform both data retrieval (read actions) and campaign execution (write actions). The platform supports the following core capabilities via natural language commands:

Retrieve Analytics: Pull performance data, look up campaign metrics, and track engagement insights
Manage Campaigns: Create entirely new omnichannel campaigns or update existing configurations
Build Audiences: Build signal-based target groups and define specific account segments
Execute Plays: Set up, refine, and modify active omnichannel sequences

AI Agents

Propensity AI allows you to create custom agents with specific functions tailored to your operational needs. You can build these specialized agents to assist with unique GTM tasks.

To build a custom agent, you describe your required role and responsibilities directly within the platform. Propensity AI then guides you through the creation process to configure and deploy your custom agent.

Read vs. Write Actions

To ensure platform security and data governance, Propensity AI categorizes user requests into distinct action levels:

Read Actions (Data Retrieval)

Read actions are safe, low-risk queries that only extract information from your workspace. The AI can pull engagement metrics, review audience criteria, track warmth score shifts, and generate pipeline summaries.

Write Actions (Platform Modification)

Write actions alter your data or live infrastructure. This includes creating new campaigns, editing active Plays, modifying segmentation logic, adjusting live budgets, or publishing newly generated creative assets. Available actions depend directly on your workspace permissions and user role configuration.

The ABM Coach

The ABM Coach allows you to identify untapped market gaps and target them using automated intelligence. To use this feature, you input your company URL directly into the platform interface.

It analyzes the website to uncover your largest market opportunities. Following the analysis, the tool automatically builds customized ABM campaigns. These campaigns include a structured target audience tailored specifically to your product or service.

Natural Language Prompt Patterns

Propensity AI requires clear and specific parameters to execute write actions effectively.

Good Prompt (Explicit Criteria): "Build an audience of enterprise SaaS accounts in North America researching cloud security in the last 30 days."
Ambiguous Prompt (Broad Execution): "Launch a campaign for security buyers." (The platform typically generates a very general campaign because it lacks specific criteria, which is not ideal for targeted ABM execution)