Student Voice AI provides a deterministic machine‑learning platform that automatically ingests, classifies, and scores sentiment for every free‑text comment from UK higher‑education surveys, using a taxonomy trained on hundreds of thousands of student responses. The system delivers sentence‑level, fully auditable analyses with sector benchmarking across 100+ institutions, producing exportable insight packs and dashboards for quality‑enhancement, accreditation and board reporting.
Funding
Funding not disclosed
Founders
Product
Problem
UK universities collect thousands of free‑text comments each year from surveys such as the NSS, PTES, PRES and module evaluations, but they lack tools that can analyse every comment quickly, reproducibly, and with higher‑education‑specific taxonomy, leaving valuable student insight unused for decision‑making.
Solution
Student Voice AI offers a deterministic machine‑learning platform that automatically ingests, classifies, scores sentiment and benchmarks all free‑text comments from UK higher‑education surveys. The system uses a taxonomy trained on hundreds of thousands of hand‑labelled student responses, ensuring categories reflect how students talk about teaching, assessment, support and other university services. Analyses are performed at the sentence level, providing granular attribution of mixed comments and complete coverage without sampling. Results are versioned, auditable and include sector benchmarks from over 100 UK institutions, enabling universities to compare their performance and produce TEF‑ready narrative evidence. Outputs are delivered as executive summaries, thematic deep‑dives, equity‑focused views and dashboards that can be directly incorporated into board papers, quality‑enhancement plans and accreditation submissions.
Target Audience
Primary customers are university quality‑enhancement teams, student‑experience offices and senior academic leaders responsible for NSS, PTES, PRES and other survey reporting, as well as institutional research analysts who need reproducible, benchmarked qualitative evidence.
Features
- Deterministic ML classification (same input always yields identical output) with no reliance on public LLM APIs
- Full‑comment, sentence‑level analysis covering every valid response across NSS, PTES, PRES, UKES and module evaluations
- HE‑specific taxonomy and sentiment model trained on hundreds of thousands of UK student comments
- Sector benchmarking against 100+ UK institutions via AdvanceHE partnership, with over‑/under‑index visualisations
- Versioned runs and complete audit trails to satisfy TEF panel scrutiny and governance requirements
- Secure UK/EU data residency, ICO‑aligned processing and intelligent redaction of distressing content
- Exportable insight packs, dashboards and TEF‑style narrative reports ready for board and quality‑enhancement use