✦ AI Learning Intelligence

Close the gap between
disengagement and institutional response

SomaAI predicts student attrition risk before it becomes dropout. Personalised paths, real-time risk signals, and content efficiency analytics — across every tenant, at scale.

Request a DemoExplore the Platform →
23%
Average reduction in student dropout rates
45+
Institutions across Sub-Saharan Africa
2.1×
Improvement in course completion rates
50k
Active learners across all deployments
The Problem

Students disengage weeks before institutions notice

The average time between a student's first disengagement signal and an academic advisor's response is 43 days. By then, the learner is already gone.

SomaAI monitors 127 behavioural signals continuously — login patterns, content consumption, assessment performance, peer interaction — and surfaces risk to instructors and advisors the moment patterns shift. Not 43 days later. At Week 1.

Student Attrition Timeline
Typical journey from disengagement to dropout — without SomaAI
Week 1 — First signal
Student misses a session; login frequency drops 40%
Week 3 — Pattern emerging
Assessment scores declining; discussion participation stops
Week 6 — Institutional notice
Advisor notified via manual report — two-week processing lag
!
Week 7 — Student withdraws
43 days after the first signal. Intervention was possible.
With SomaAI — Week 1 alert
Advisor notified at Week 1. Personalised outreach by Week 2. Student retained.
What SomaAI does

Built for the entire learning lifecycle

From the first login to certificate completion, SomaAI provides the intelligence layer that connects learner behaviour to institutional action.

AI Risk Prediction

Graph Attention Networks trained on 3 million student journeys surface disengagement risk in real time, per student, per course — with explainable reasoning.

  • 127 behavioural signals tracked
  • 14-day early warning
  • 78% AUC model accuracy
  • FNR gap < 5pp across demographics
🗺

Personalised Learning Paths

When assessment reveals a knowledge gap, SomaAI generates a remedial path in seconds — specific modules, in the right sequence, for that learner.

  • Adaptive to each learner's pace
  • Remedial path generation
  • Skill gap analysis
  • Confidence-scored recommendations
📊

Multi-Tenant Analytics

Institution-wide dashboards break down completion rates, engagement scores, and risk distribution by cohort, course, instructor, and demographic stratum.

  • Isolated per institution
  • Custom branding support
  • Role-based access control
  • GDPR-compliant data handling
🎯

Content Efficiency Engine

Identify which modules drive completion and which cause drop-off. Data-driven recommendations for content teams — at the item level, not the course level.

  • Content-item level analytics
  • Engagement drop-off detection
  • A/B content comparison
  • Auto-flag low-efficacy content
Deployment

Live in 3 weeks. No rip-and-replace.

SomaAI sits alongside your existing LMS and begins generating risk signals from day one. No new infrastructure. No migration.

1

Connect your data

LTI 1.3 and xAPI connectors for Moodle, Canvas, Blackboard, or a custom LMS. Historical data ingested in the first 48 hours.

2

Calibrate the model

Our team calibrates the GNN on your institution's historical outcomes. Fairness gates validated across your student demographics before go-live.

3

Activate and monitor

Real-time risk signals, advisor dashboards, and automated learner nudges go live. Weekly outcome reports delivered from Week 1.

What educators say

Trusted by educators across Africa

Within one semester, our academic support team was intervening three weeks earlier on average. Foundation Year retention went from 68% to 84%. SomaAI is the only tool that tells us why a student is at risk — not just that they are.

Dr. Diana Mwangi
Dean of Academic Affairs, Strathmore University · Nairobi

Start reducing preventable attrition today

30-minute demo · No commitment · Deployment feasibility assessment included

Schedule a DemoSee the Learner View →