The goal is for Salesforce to recommend optimal discount, target price, or margin when users create or edit a Quote Line Item — based on customer profile, historical deal performance, competitor pricing, and product margins.
You will have a Prediction Service that produces a recommended price or discount and exposes it in Salesforce UI.
Amount
Probability
StageName
CloseDate
Industry (Account)
Type
Forecast Category
Account
AnnualRevenue
Industry
BillingCountry
No of Past wins / losses
ListPrice
Cost (standard Cost__c or StandardUnitCost in Product2)
Product Family
Product Category
Quantity
Discount (%)
Net Total
Competitor__c (custom but non-sensitive)
Deal Size bucket field
Margin (formula: (ListPrice - DiscountedPrice) - Cost)
2. AI Model Design
Einstein Discovery (Native)
Train a model on all closed-won and closed-lost deals + historical pricing data.
Prediction objective:
“Optimize Recommended Discount to maximize win probability while protecting margin ≥ X%.”
Outputs:
Recommended Discount
Reason Codes ("High win rate for similar customers")
Confidence Score
On Quote Line Item save → trigger → Apex → Named Credential → AI endpoint → return pricing recommendation.
A screen flow in Quote Builder:
User selects product + quantity.
Flow calls Apex Action → AI Model → Returns Recommended Price.
Shows recommended price in modal.
User accepts or overrides.
Use the Einstein Component on Quote Line Item page → automatically shows recommendations.
"Get AI Pricing Recommendation" button Shows:
Recommended Discount
Recommended Price
Confidence Score
Reason Codes
Maximum Discount Allowed per Product Family
Minimum Margin Allowed
Competitor-specific pricing rules