Corpus reuse in the deployed support system is concentrated on a single process step: three certificate/offer-letter entries account for 56.8% of 871 resolutions, with a Gini of 0.586 across entries¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Reuse of the 114-entry FAQ corpus is highly concentrated: entries 4, 3 and 1, all concerning certificate submission and offer letter issuance, account for 56.8% of the 871 corpus-mediated resolutions, and the Gini coefficient of reuse across the 113 well-formed entries is 0.586. → Gupta 2026

Evidence¶

Gupta 2026¶

Gupta, J., Ayinampudi, P., Aditya B.M.V., Hegade, P., Sharma, R., Sharma, S., Meenakshi, V., Iyengar, S.R.S. (2026). What Students Actually Ask: Demand Structure and Automation Potential in a Hybrid Support System. arXiv:2609.38967. https://arxiv.org/abs/2609.38967 doi:10.48550/arxiv.2609.38967

q2 · i? · design · r2

Descriptive analysis of knowledge-reuse volume across corpus entries in the nine-week operational record (N=871 corpus-linked queries). The three top entries cover "certificate submission and offer letter issuance"; median reuse per entry is 3 against a mean of 7.7, and 30 entries were used exactly once.

"Entries4,3and1,allconcerningcertificatesubmission andofferletterissuance, account for 495 of 871 resolutions, or 56.8%, and no other cluster exceeds 9.0%. Acrossthe113wellformedentriestheGinicoefficientofreuseis0.586"

Discussion¶

Related Claims¶

• AI agent communities exhibit extreme participation inequality (comment Gini = 0.91), exceeding human online learning communities — related
• Support demand across participants is concentrated: the most active decile accounts for 34.6% of queries and 42.9% of participants raised exactly one query — related
• Operational resolution records require workflow-semantic checks before interpretation: one administrator accounts for 72.7% of attributed human resolutions, most administrative closures occur in synchronous bulk events, and the accepted_derived field encodes resolver identity rather than acceptance — related
• At least 20.4% of queries in an online internship programme are status enquiries about pending submissions, which a corpus-grounded AI assistant served only 1.3% of the time — related
• Only 8.0% of queries reaching a person duplicated existing corpus content, suggesting the knowledge base was already well used — related
• The axiom dataset is work-local: 1,448 of 1,455 axioms map to exactly one work — related