A University Faculty Manages 30,000 Theses

"Alumni used to wait three weeks for a transcript. Now: 48 hours." — Faculty Registrar

The numbers

30,000+ postgraduate theses ingested
3 decades of historical archive digitised
48 hours turnaround on alumni transcript requests (down from 3 weeks)
94% thesis classification accuracy after fine-tuning

The scope

A Faculty of Postgraduate Studies at one of Kenya's largest public universities. Theses spanned 30+ years across:

Masters dissertations
PhD theses
Postdoctoral research reports
Examiner reports
Defence panel records

The physical archive occupied two rooms; the digital archive lived on a network drive with inconsistent metadata.

The challenge

Three audiences needed access:

Current students researching prior work (couldn't find it; resorted to Google Scholar even for in-house work)
Faculty building bibliographies (similarly struggled)
Alumni requesting copies of their own theses for further academic / professional use (3-week turnaround)

Plus an organisational requirement: CUE accreditation reviews ask for thesis-related records and would not accept the existing archive as compliant.

The migration

A two-phase digitisation + ingestion programme over 14 months:

Phase 1 — Born-digital theses (months 1-3): The last 8 years of theses already existed in PDF on the network drive. Ingested directly, metadata extracted via AI (title, author, supervisor, year, abstract, keywords).

Phase 2 — Historical scanning (months 4-14): Outsourced scanning bureau processed the physical archive. Roughly 22,000 historical theses scanned at 300 DPI, OCR'd, ingested. Spot-checking at 2% sampling rate confirmed quality.

The outcomes

Student research: Theses now appear in semantic search by topic, not just by author/year. A student researching “drought-resistant maize hybrids” surfaces relevant historical work without remembering specific titles.
Faculty bibliographies: Plagiarism checks against the in-house corpus added a layer beyond commercial tools.
Alumni transcripts: Transcript requests now follow a Papyrus workflow — identity verification, registrar approval, signed PDF delivery via secure share link. Turnaround averages 48 hours.
CUE accreditation: The next review accepted the digital archive as compliant. The dean's response: “Finally.”

What stalled

Some faculty members were uncomfortable with their early-career work becoming easily searchable. The faculty addressed this with a clear access policy:

Public: thesis title, author, abstract (always)
Faculty-internal: full thesis content
External (alumni / institutional): on request via the transcript workflow

This is the policy that already governed physical access; the digital version simply enforced it consistently.

Quote

"The 1996 thesis cited in last year's PhD defence — that was found in Papyrus. Twenty-seven years on a shelf, then suddenly relevant." — Faculty Dean