Research guide
How the research guide works
How an answer is produced
When you send a question, it goes to this site's own server, which adds a written summary of the published research themes, case studies, projects and collaboration areas before passing your question to the model. The answer is generated live for your question — it is not selected from a list of prepared replies, and no two answers will be identical. Responses arrive a sentence at a time as the model writes them.
The instruction is deliberately narrow: recommend the research themes, case studies and collaboration areas that best match what you describe, link to the pages that hold them, and be explicit when the published work does not go far enough.
What it is grounded in
Research themes
The five themes published on the research page: human behaviour and decision-making, digital trust, AI in healthcare, human-centred AI, and policy and ethics.
ResearchCase studies
The three portfolio case studies — Algorithms of Trust, digital trust in healthcare decision-making, and the Ri Vha Venda leadership work — including their context, approach and outcome.
PortfolioProjects and collaboration areas
The current research programme and the areas in which Ndugelo works with organisations, as described on the projects and consulting pages.
ConsultingWhat it will not do
- It is not Ndugelo. It cannot speak for him, commit him to work, negotiate terms, or confirm availability.
- It only knows what is published on this site. It has no access to unpublished drafts, dissertation data, interview transcripts, or private correspondence.
- It does not search the web or cite outside literature, so it cannot summarise other researchers' work or current news.
- Where the published material does not answer a question, it is instructed to say so and point you to the contact page rather than guess.
- Like any language model it can be imprecise. Treat anything consequential as a starting point and confirm it with Ndugelo directly.
Your conversation
Conversations are kept in your own browser so you can return to them, and are not shown to anyone else. Clearing your browser storage removes them. Your question is sent to the model in order to answer it; nothing is added to a mailing list, and you are never asked to identify yourself.