Knowledge Base
GPTWeb Engagement Scoring: How It Works
GPTWeb's scoring engine goes well beyond simple page-view counting. It uses AI to evaluate the quality and depth of every visitor interaction — meaning a visitor who has two rich, intent-filled conversations can outscore someone who visited 20 times with shallow engagement. Scores run 0–100, update continuously, and feed directly into DQL qualification, AI Segmentation, and Agentic Workflows automatically.
0 – 100
Score Range
70+
Default DQL Threshold
4 Dimensions
Scoring Factors
Auto-Aggregated
Account Scoring
The 4 Scoring Dimensions
Scoring Factor Breakdown
| Factor | Weight | What It Tracks | Example Signal |
|---|---|---|---|
| Engagement | 25% | Chat activity volume, response quality, conversation depth and richness | A visitor asking detailed product questions vs. a one-word reply |
| Behavioral | 25% | Activity patterns, visit frequency, session engagement duration | Returning 3x in a week with long sessions vs. single short visit |
| Firmographic | 25% | Company size, industry vertical, estimated revenue | Enterprise company in target vertical scores higher |
| Demographic | 25% | Job title, department, seniority level | VP of Marketing scores higher than an intern |
Key Scoring Intelligence Signals
Beyond the four core dimensions, the AI scoring engine tracks specific intent signals that dramatically influence the final score:
Intent Signals Tracked by the AI Scoring Engine
| Signal | Weight | What It Means | Scoring Dimension |
|---|---|---|---|
| Pricing Questions | High | Visitor asks about cost, plans, or packages — strong buying intent | Behavioral + Engagement |
| Demo Requests | High | Visitor initiates or clicks demo scheduling links | Behavioral + Engagement |
| Enterprise Feature Interest | High | Visitor asks about SSO, API, admin controls, or scale | Demographic + Engagement |
| Topic Diversity | Medium | Breadth of subjects covered across conversations — signals serious evaluation | Engagement |
| Return Visit Frequency | Medium | Repeat visits with active conversations — not just passive browsing | Behavioral |
| Form Completions | Medium | Submitted contact, download, or campaign forms | Behavioral |
| Conversation Depth | High | Long exchanges with substantive back-and-forth vs. surface questions | Engagement |
| Time on Site | Medium | Duration of active sessions — weighted by engagement quality | Behavioral |
Account-Level Scoring
GPTWeb automatically aggregates individual visitor scores up to the account (company) level — no configuration required. The Account Score modal shows: - Aggregate score across all contacts at that company
- Contributing contacts and their individual scores
- Derived Personas for each contact
- Continuous recalculation as new activity arrives This is the backbone of Intelligent ABM — understanding account-level intent, not just individual visitor behavior.
Scoring to DQL Pipeline
Configuring Your Scoring Model
Scoring Configuration Options
| Setting | Location | Purpose |
|---|---|---|
| Scoring Model | Admin Panel → AI Scoring | Choose which factors to emphasize for your audience |
| Score Weights | Admin Panel → AI Scoring → Weights | Adjust the 25/25/25/25 default split per your ICP |
| DQL Threshold | Admin Panel → AI Scoring → Thresholds | Set the score (default 70) that triggers DQL status |
| Alert Threshold | Admin Panel → AI Scoring → Thresholds | Set when to fire sales notifications |
| Decay Rate | Admin Panel → AI Scoring → Settings | How quickly old inactive scores fade over time |
| Recalculation Frequency | Admin Panel → AI Scoring → Settings | How often scores update as new activity arrives |
One powerful downstream effect of scoring: as visitor scores change, segment membership updates automatically. Combine this with AI Campaigns and you have a fully automated nurture loop — visitors score up, move segments, trigger agents, and receive perfectly timed outreach without any manual intervention. That's AI Marketing Automation working as it should.
GPTWeb is the future of engagement — websites and marketing automation combined, built for the AI era, built for now.