AI Outbound Scoring: Building Scoring Models

Audience: Admin (typically RevOps)

AI Outbound Scoring lets you build the scoring models that power prospecting in Sales Copilot. Each model evaluates companies across three dimensions: Fit, Need, and Intent. Models draw on the entire HG dataset and take only a few minutes to create.

How scoring models relate to offerings

Scoring models are built per offering and associated with that offering. Because each offering carries its own ICP, pain points, intent topics, and GTM motions, its scoring model reflects that specific buying motion rather than a one-size-fits-all score. This is what allows sellers to run concurrent plays across different offerings, each with its own signal prioritization.

Creating a model

From Scoring, select Create model. The AI Assistant walks you through three steps:

1. Define company

2. Build models

3. Review & save

Step 1: Define company

Confirm your company domain. Then choose what anchors the model:

• Select an offering to build a model around that offering's context and GTM motions (such as net-new-logo, displacement, or expansion), or

• No offering to build from the domain alone.

You can also optionally tell the AI Assistant what you want the model to focus on. It uses this, along with your company context, to research your business and build scoring models for you to review.

Steps 2 and 3: Build and review

The AI Assistant builds the model, then presents it for your review before you save it. Once saved, the model is used to score accounts for prospecting.

Why Fit, Need, and Intent

Each dimension answers a different prospecting question:

• Fit: does this company look like your ICP?

• Need: does this company have the pain your offering solves?

• Intent: is this company showing in-market signals right now?

Scoring across all three, grounded in HG data, gives sellers a prioritized view of which accounts are worth their time and why