Proactive Approach to Student Retention

Inspiration

Student Success Advisors in higher education institutions often struggle to proactively identify and support at-risk students due to fragmented systems and siloed data. The idea for this project stemmed from the need for a centralized, intelligent system that could unify student data and enable timely, personalized interventions to improve retention and academic performance.

What it does

The solution, "Proactive Student Retention Using Agentforce," leverages Salesforce technologies to:

• Unify and harmonize student data from SIS, LMS, Financial Aid, and other systems using Salesforce Data Cloud.
• Train Predictive AI models via Einstein Studio to assign at-risk scores based on engagement patterns, GPA, attendance, etc.
• Query these AI models via Education Cloud to highlight contributing factors and suggest targeted interventions.
• Use Generative AI (via Agentforce) to convert model outputs into human-readable insights, allowing advisors to take timely action.

The output includes:

• At-Risk Score
  
• Contributing Factors
  
• Recommended Interventions

How we built it

We built the solution using the following tech stack:

• Salesforce Data Cloud for data ingestion, cleaning, harmonization, and transformation.
• Einstein Studio for training machine learning models to assign risk scores.
• Education Cloud to visualize and operationalize risk insights.
• Agentforce with Generative AI to convert model outputs into student-specific recommendations.
• MuleSoft Anypoint Platform to integrate SIS, LMS, and CRM systems into the Data Cloud.

The process involved:

1. Data ingestion and unification using CDP models.
2. Model development with preprocessing, training, and validation steps.
3. Score querying through Education Cloud.
4. Generative insights via Agentforce for advisor action.

Challenges we ran into

• Ensuring data harmonization across disparate systems (SIS, LMS, CRM).
• Managing model accuracy while using diverse datasets with varying data quality.
• Designing a clear and actionable output from the Predictive + Generative AI pipeline.
• Maintaining compliance and data privacy during student data processing.

Accomplishments that we're proud of

• Created an end-to-end AI-driven intervention pipeline from ingestion to advisor insight.
• Built a working prototype that can easily scale across educational institutions.
• Successfully integrated multiple Salesforce clouds and Agentforce for real-time insight delivery.
• Developed a human-readable, actionable output that directly empowers Student Success teams.

What we learned

• The power of harmonized data cannot be overstated—unified insights drive better decisions.
• Predictive models must be contextualized through human-readable formats to be actionable.
• Collaboration between AI, cloud platforms, and human support staff leads to meaningful change.
• Generative AI can bridge the gap between technical model output and real-world educational interventions.

What's next for Proactive Approach to Student Retention

• Scale the model for broader academic metrics, including extracurricular involvement and mental health indicators.
• Deploy a mobile-first advisor dashboard for real-time risk alerts and intervention tracking.
• Integrate WhatsApp and student communication tools for faster advisor-student engagement.
• Enhance personalization using deeper behavioral and social-emotional analytics.
• Build a feedback loop to continuously improve model predictions based on intervention outcomes.

Built With

• agentforce
• datacloud
• education
• educationcloud
• einsteinstudio
• genai
• predictiveai