Platform Capabilities

The complete AI stack for student success

From enrollment to graduation — SomaAI predicts risk, personalises paths, and closes the attainment gap before it opens.

Request a DemoView Pricing →
What SomaAI Does

Six capabilities. One platform.

Every feature is purpose-built around a single goal: closing the information gap between learner disengagement and institutional response.

AI Risk Prediction

Identifies struggling learners 14 days before failure

Graph attention networks trained across EdNet, MoocCubeX, and OULAD datasets. 78% test AUC. Identifies at-risk learners 4–6 weeks before dropout — with explainable per-signal reasoning for every instructor.

🗺

Personalised Learning Paths

Adaptive paths generated from performance data

AI generates targeted remedial paths in under 90 seconds when a learner underperforms. Adapts to pace, learning style, and engagement history. Goal-based, remedial, and skill-gap path types all supported.

🏛

Multi-Tenant Architecture

Isolated environments for each institution

Serve 12+ institutions from a single platform with full namespace isolation, per-tenant branding, and 4-tier role access (learner, instructor, admin, super-admin). No data bleed between tenants.

📊

Real-Time Content Analytics

Track engagement at the content-item level

Per-module engagement rates, difficulty scores, and completion trends updated every 15 minutes. No separate BI tooling needed. Identify bottleneck items before they become dropout accelerators.

Fairness-First AI

Bias-corrected models with confidence scores

FNR gap monitoring keeps false-negative rate differences below 5 percentage points across demographic strata. Bias detection runs on every model retrain. Quarterly fairness audit reports are published publicly.

Assessment Suite

Auto-grading with rubric-based feedback

Multiple-choice, short answer, file upload, peer review, and proctoring-ready tracking with per-question item analysis. Attempt history, score trends, and mastery tracking per learner per skill.

How It Works

From signal to intervention in three steps

SomaAI is designed around a closed loop: ingest data, surface risk, trigger action. Each step is automated; each alert is explainable.

Signal
Analysis
Intervention
1

Ingest & Baseline

Connect your existing courses and learner data via LTI 1.3 or xAPI. SomaAI establishes risk baselines and cohort profiles within 24 hours of connection.

2

Monitor & Predict

The GAT model scores every learner daily across 127 behavioural signals. Instructors see at-risk flags weeks before they become dropouts — with the exact signals that drove each score.

3

Intervene & Improve

Personalised paths, targeted nudges, and escalation alerts close the attainment gap. Real outcomes tracked per cohort. Weekly improvement reports delivered automatically.

78%
Test AUC — risk detection accuracy
4–6
Weeks early warning before dropout
12
Institutions currently on the platform
<90s
Personalised path generation time

SomaAI cut our second-semester dropout rate by 23% in the first cohort. The at-risk dashboard alone saved our intervention team 40 hours a week — time they now spend actually talking to students.

AB
Dr. Ama Boateng
Head of Digital Learning, University of Ghana

Ready to close your attainment gap?

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

Schedule a DemoView Pricing →