About Student Voice AI
Student Voice AI builds deterministic machine learning tools that turn student survey comments into decision-grade evidence for UK universities. Founded by Dr Stuart Grey SFHEA, we started as a university spin-out driven by a simple observation: institutions collect thousands of free-text comments every year but lack the tools, time, and methodology to analyse them rigorously at scale.
Our founder
Dr Stuart Grey is a Senior Fellow of the Higher Education Academy (SFHEA) and an active lecturer at a Russell Group university. Having spent years teaching and working within HE institutions, he saw first-hand how student feedback could transform teaching and learning, but how traditional methods of analysis were slow, inconsistent, and limited to small samples. This gap between what institutions collected and what they could act on inspired the creation of Student Voice Analytics.
Stuart's dual role as a practising academic and company founder means the platform is built with a deep understanding of institutional realities: TEF submission deadlines, Board reporting cycles, Quality Enhancement frameworks, and the day-to-day pressures facing survey leads and student experience teams.
What we do
Student Voice Analytics classifies, scores sentiment, benchmarks, and summarises student comments from NSS, PTES, PRES, UKES, module evaluations, and custom surveys. Unlike generic text analytics or LLM-based tools, our models are deterministic: run the same comments twice, get the same results. Every output is versioned, auditable, and traceable from summary insight back to source comment.
Our HE-specific taxonomy was trained on hundreds of thousands of hand-labelled student comments from across the UK sector. This means the models understand how students talk about their education, not how consumers talk about products or services.
Our methodology
- Deterministic ML, not LLM: Reproducible classification that withstands TEF panel scrutiny. Same data in, same results out, every time.
- All-comment coverage: Every comment analysed at sentence level. No sampling, no blind spots, no mixed-sentiment comments miscategorised.
- Sector benchmarks: Compare your institution's feedback patterns against 100+ UK HE institutions via our AdvanceHE partnership.
- Governance-ready outputs: Versioned analysis runs, full audit trails, and TEF-ready documentation as standard.
Trusted by leading institutions
We work with universities across the UK, from Russell Group research-intensives to post-92 institutions and specialist providers. Our institutional partners include the University of Edinburgh, University of Leeds, King's College London, University of Warwick, UCL, The Open University, Queen's University Belfast, and the London School of Economics.
Through our partnership with AdvanceHE, our sector benchmarks draw on data from over 100 UK HE institutions, giving teams the context they need to understand whether feedback patterns are institution-specific or sector-wide.
Data protection
All analysis runs on our own infrastructure with UK/EU data residency. We do not use public LLM APIs to process student data. Our data processing is ICO-aligned, and we provide full data processing agreements for institutional procurement. Student wellbeing is protected through intelligent redaction of distressing content.
By higher education, for higher education
Our philosophy is simple: the people who build tools for universities should understand universities. Stuart still teaches. Our taxonomy was built with HE professionals. Our roadmap is shaped by conversations with PVCs, Directors of Quality, and survey leads across the sector. We are not a generic text analytics company that pivoted into education.
