Salesforce Next Best Action: AI-Powered Recommendations at Every Customer Touchpoint
8 capabilities that surface the right action for every agent, at every moment, in every channel.
Service Agents Who Cannot Upsell. Sales Reps Who Cannot Personalize.
Most sales and service teams miss revenue and retention opportunities because the right recommendation is buried in systems the agent cannot access in real time, or because the agent does not know what the right recommendation is at all.
- Service agents resolving cases without knowing what the resolved customer could benefit from next, upsell moment missed
- Sales reps delivering the same pitch regardless of the buyer's history, sentiment, or channel
- Marketing offers appearing in service conversations that are inappropriate for the customer's situation or regulatory status
- No data on which recommendation strategies actually convert, managers operating on intuition rather than evidence
- Recommended actions not adapted to the channel: a chat conversation needs a different action than a phone call
Eight NBA Capabilities That Make Every Agent Smarter in Real Time
Salesforce Next Best Action embeds a context-aware recommendation engine directly into the agent's Salesforce screen, surfacing the right action at the right moment without requiring the agent to ask for it.
- Einstein AI recommendation engine embedded in Service Cloud and Sales Cloud: surfaces top 3 recommended actions on every case, opportunity, and contact screen
- Context-aware: reads case history, purchase history, last interaction sentiment, and current channel before generating recommendations
- Upsell and cross-sell recommendations surfaced at case resolution, the highest-intent moment in a service interaction
- Channel-appropriate actions: phone call recommendations emphasize verbal prompts; chat recommendations emphasize quick-link offers; email recommendations surface longer-form content
- Built-in A/B testing framework: compare two recommendation strategies in live production with automatic winner selection based on conversion rate
- Compliance filter: automatically excludes any recommendation that the customer is not eligible for under regulatory rules or account restrictions
- Manager override: custom recommendation rules per product line, business unit, or customer segment configured by revenue operations without code
- ROI attribution: every NBA action tracked with conversion rate and revenue attribution, the first time most organizations see the actual value of their recommendation engine
What Next Best Action Delivers Across the Business
NBA changes the economics of every customer-facing interaction by turning each touchpoint into a data-driven decision.
- Service interactions generate revenue opportunities instead of just resolving issues
- Recommendation acceptance rates improve as Einstein AI learns which actions convert for each segment
- Compliance risk eliminated through automated product eligibility filtering
- A/B testing replaces intuition-based recommendation strategy decisions with evidence
- Revenue operations can tune recommendation strategies in minutes without engineering involvement
- ROI attribution shows the exact revenue impact of the NBA program for the first time
Salesforce and Agentforce Components Used
Common Questions About This Deployment
What is the difference between Salesforce Next Best Action and a standard recommendation engine?
Standard recommendation engines typically operate on a single signal (purchase history) and return a static result. Salesforce NBA reads multiple real-time signals simultaneously, case context, sentiment, account tier, current channel, regulatory status, and applies business rules on top of the Einstein AI model to surface a recommendation that is not just data-optimal but also business-appropriate and compliant.
How does the A/B testing framework work in production?
Strategy Builder randomly assigns each customer interaction to one of the configured strategy variants. Both strategies run simultaneously with the same customer population. Einstein Analytics tracks conversion rates per variant and surfaces statistical significance when a winner emerges. The revenue operations team activates the winning strategy with a single configuration change, no code deployment required.
Can NBA recommendations be triggered outside of agent interactions, for example, in automated email journeys?
Yes. Salesforce NBA can integrate with Marketing Cloud Journey Builder to surface NBA recommendations in automated email and SMS sequences. The recommendation engine runs at send time, not at journey creation time, so each recipient receives a recommendation based on their context at the moment the message is sent, not a static segment recommendation built weeks earlier.
How long does it take for the Einstein AI model to learn and improve recommendations?
The initial model is configured with business rules and initial strategy definitions from day one. The learning loop begins as soon as interaction and outcome data accumulates. Most clients see the model's recommendation quality materially improve within 60-90 days as the AI identifies which actions are actually converting for each customer segment in their specific context.
More Capability Spotlights
Build this outcome in your organization.
Every Agentforce engagement starts with a free discovery call to map your use case, validate feasibility, and scope a fixed-price delivery. The first 2 weeks carry zero risk.
