Student Success & Retention Solution

The Vendor is required to provide to modernize its student success infrastructure to strengthen retention, enhance student satisfaction, and deliver a more coordinated, data-driven approach to supporting students throughout their academic journey.
- This initiative aligns with the university’s strategic priorities to increase enrollment to 23,000 students, achieve a 90% retention rate, and improve academic achievement, career preparedness, engagement, and holistic wellbeing.
- Provide to procure a comprehensive Student Success and Retention (“SSR”) platform that enables early identification students at-risk of not persisting, monitors academic progress, and supports holistic student wellbeing.
- The desired solution will integrate multidimensional student data, automate workflows, and facilitate targeted, timely interventions.
- The goal is to move from reactive data and support to proactive predictive analytics and automated workflows. By unifying communication and care coordination across departments, the platform will empower faculty, advisors, and student support professionals to collaborate effectively and ensure that every student has a clear pathway to success.
- Predictive Analytics Engine
• Analyze historical and continuous, low-latency, bi-directional data to identify student success factors and at‑risk persistence indicators.
• Generate individualized risk scores and propensity models for retention, course success, and term‑to‑term persistence.
• Provide transparent model explanations to support sustained faculty engagement, advising case management, campus partners utilization, and student success and retention understanding and intervention planning.
• Continuously refine models using machine learning and updated institutional data.
• Provide documentation of all input variables used in predictive models.
• Identify minimum data completeness thresholds required for a valid model performance.
• Describe how missing or delayed data affects risk scoring outputs.
• Demonstration of bias testing methodologies and equity impact monitoring.
• Provide validation of metrics based on institutions with similar characteristics.
• Describe risk model recalibration frequency.
• Describe drift detection methodologies.
• Describe their processes and practices of institutional access to performance metrics by subpopulation.
• Describe implemented predictive algorithms.
• Explain how model and algorithm bias and overfitting are discovered and addressed.
- Machine Learning Modeling
1. Explain ability and Transparency – The platform shall:
• Describe machine learning technologies.
• Provide clear, interpretable insights into why it identifies a student as at-risk.
• Transparent model explanations - detail how explainable AI is implemented in the platform, e.g. by showing key predictors or reasons for each risk alert, rather than just a mysterious score.
• A well-designed AI-powered system should augment human decision-making, not replace it.
• Explain how model and algorithm bias and overfitting are discovered and addressed.
2. Customization & Maintenance
• Models must be configurable to agency context and allow adjustment of variables or weighting.
• Vendors should explain how models are updated or retrained over time and how accuracy is maintained.
3. Evidence of Effectiveness
• Provide clarity on what the AI actually does (predictive modeling, alerts, chatbot support, etc.).
• Provide supporting evidence such as case studies or measurable impact from comparable institutions.
4. Data Use
• If the platform’s AI models are trained using our student data - agency retains ownership of our data and any derivative models.
• Machine learning models trained on agency data cannot be used or shared externally without permission, or that an institution-specific model option is available for privacy.