How AI Deal Scoring Works in Dalil

Dalil AI automatically scores every deal in your pipeline from 0–100, telling you which deals to focus on, which are at risk, and which are about to close.

How the Score Is Calculated

The AI combines three categories of signals:

1. Engagement Signals (40% weight)

• Email opens, clicks, and replies
• Time spent on proposals and documents
• Website visits (via tracking pixel)
• Calendar link clicks and meeting acceptance rate

2. Behavioral Signals (35% weight)

• Days since last meaningful interaction
• Number of stakeholders engaged in the account
• Response time patterns (are they getting slower or faster?)
• Stage progression velocity vs. your team's historical average

3. Fit Signals (25% weight)

• Company size and industry match to your ICP
• Tech stack alignment (via enrichment)
• Title and seniority match to your typical buyer
• Funding stage and growth trajectory

Score Ranges and What They Mean

Score	Status	Recommended Action
85–100	🟢 Hot — closing soon	Call within 24h, prepare contract
65–84	🟡 Progressing	Next step defined, follow-up on schedule
40–64	🟠 Needs attention	Re-engage, add new stakeholder, change angle
0–39	🔴 At risk	Assess if still active, consider closing as lost

Training the AI on Your Data

Out of the box, the AI uses industry benchmarks. After 90 days, it trains on your team's specific win/loss data and becomes significantly more accurate.

To speed this up:

1. Go to Settings → AI → Deal Scoring
2. Click Train on historical data
3. Select your last 12 months of closed deals
4. The AI analyzes patterns and updates its model (takes ~2 hours)

Score Alerts and Notifications

Set up Slack or email alerts when deals change score significantly:

1. Go to Settings → Notifications → Deal Score Alerts
2. Set thresholds: e.g., "Alert me when any deal drops more than 20 points in 48 hours"
3. Choose delivery: Slack DM, email digest, or both

Overriding a Score

You can manually adjust a deal's score and tell the AI why. This helps it learn your judgment:

1. Click any deal → Score badge → Adjust score
2. Enter your score and select a reason
3. The AI incorporates this feedback going forward

FAQ

Can the AI score be wrong? Yes — it's probabilistic, not deterministic. Think of it as a prioritization signal, not a definitive prediction. Your judgment always takes precedence.

Does it work for different sales motions? Yes — the AI adapts to PLG, outbound, inbound, and enterprise motions. You specify your motion during onboarding.

What if I have very few deals? Below 10 active deals, the AI uses pure industry benchmarks. As you add more data, it personalizes.