Support-query records from an internship programme's helpdesk: 4,093 queries, each with its topic theme (16 categories), how it was raised, and how it was resolved (6 pathways, such as admin, peer, AI-agent or self-resolved). It has no free-text query content, and all people are pseudonymised.
Support-query records from an internship program's helpdesk, covering how each query was raised, what it was about, and how it was resolved. The dataset has 4,093 rows, one per query, with 16 columns. It has no free-text query content. Each query's topic is given as one of 16 theme codes. Each row also has a resolution pathway, one of 6 categories, such as admin-resolved, peer-resolved, AI-agent-resolved, or self-resolved. Two small lookup files describe the themes and pathways.
The dataset supports research on how support demand and resolution work in large online internship programs. Possible uses: - Analysing which topics generate the most support queries. - Comparing resolution pathways (admin, peer, AI-agent, self-resolved) and how fast each responds. - Studying resolver workload and FAQ reuse. - Building and testing models that classify queries by theme or route them to the right resolver. - Evaluating where AI agents and peer support can reduce the load on administrators.
MIT
© 2026 - Copyright AIKosh. All rights reserved.