{"resourceId":"feng-chia-ai-smart-administration-competition-20261005","versions":[{"version":"external-f16cf29ac962c7f239f49ff6898dbf6a5a79ae16d5d6e88ff3e0041a53bc2349","resource":{"id":"feng-chia-ai-smart-administration-competition-20261005","title":"Feng Chia University HR-run competition draws 79 staff teams building AI administrative tools; winner built with Microsoft 365 Copilot and Claude Code automates student-assistant scheduling","organization":"Feng Chia University, Office of Human Resources","sector":"Higher education administration / staff-built AI workflow tools","geography":"Taichung, Taiwan (private university)","publishedAt":"Feng Chia University news, October 5, 2026; winners announced September 30, 2026","publicationDate":"2026-10-05","eventDate":"2026-09-30","sourceName":"Feng Chia University News","sourceLabel":"Institutional news release (self-reported)","sourceUrl":"https://www.fcu.edu.tw/en/news/2026100501/","evidenceClass":"vendor-claim","outcomeClass":"emerging","topics":["developers-agents","knowledge-work","operating-model","accessibility-workforce"],"finding":"Feng Chia's Office of Human Resources ran its second 'AI Leadership: Smart University Administration Innovation Competition', with 79 faculty and staff teams nominated by first-level units. The Office of Academic Affairs won with a Student Assistant Intelligent Management System built with Microsoft 365 Copilot and Claude Code: it imports class schedules to avoid conflicts, generates regulation-compliant labour attendance PDFs in one click, offers a mobile app for task selection and auto-scheduling, and is said to simplify payroll reimbursement. Runners-up were an AI website quality dashboard (Secretariat) and a venue, maintenance and cost platform (campus management). The chair said several systems are already in practice and added NT$10,000 bonuses.","sledRelevance":"A concrete model for surfacing administrative AI use cases from staff themselves, with HR as sponsor, and a sign that staff are now building working tools with coding agents rather than only using chat assistants. Facilities and payroll-adjacent workflows are in scope.","evidence":"Institutional news release naming teams, tools and features. No time savings, error rates, user counts, cost or security review details are given; 'already implemented in practice' is not quantified. International context with limited transferability.","architectureImplications":"Staff-built systems using Copilot and Claude Code that touch class schedules, attendance records and payroll create shadow-application risk unless hosting, integration and maintenance are owned by IT.","governanceImplications":"A competition is an intake channel for use cases, but the release does not mention review for data protection, records compliance or long-term support before tools go live.","securityPrivacyImplications":"The winning tool handles student-assistant schedules, attendance and payroll-related records; the release is silent on access control, data storage and code review.","caveats":"Self-reported; no metrics; unclear governance of staff-built code; Taiwanese private university context.","streamIds":["campus-operations"],"roles":{"sales":"Sales framing: Feng Chia shows staff-led AI building, sponsored by HR, producing working administrative tools for scheduling, facilities and web quality. The opening is not more ideas but helping an institution turn staff prototypes into supported, reviewed services. A bounded engagement is a review of two or three staff-built tools for security, data handling and supportability. Do not claim time or cost savings; the release reports none.","engineering":"Engineering framing: tools built with Copilot and Claude Code that touch schedules, attendance and payroll need the same review as any application. Fit depends on where the code runs, how it authenticates and which systems it reads. A proof of value should inventory staff-built AI tools, check access control and data storage, and confirm who maintains each one once the original builder moves on.","delivery":"Delivery framing: competitions produce prototypes; delivery turns the best into owned services. A practical plan defines intake criteria, a security and records review, a hosting standard and named owners before campus-wide sharing. Measure adoption and process time for each scaled tool against a baseline. Risks include unsupported code, inconsistent data protection and duplicated tools across units."},"roleDetails":{"sales":{"customerProblem":"Staff are building AI tools for administrative work faster than the institution can review, host or support them.","stakeholders":["HR / staff development","CIO and application owners","Information security","Unit administrators","Payroll and facilities leads"],"discoveryQuestions":["How many staff-built AI tools are in use?","Who reviews them before they touch records?","Where is the code hosted and who maintains it?","Which prototypes do you want to scale campus-wide?"],"valueHypothesis":"A light review-and-scale path could convert staff prototypes into reusable services while lowering risk, but this source gives no outcome evidence.","potentialEngagement":"A 6–8 week review of a few staff-built AI tools with a scale-up checklist and ownership model.","unsupportedClaims":["Time or cost savings","Error reduction","That tools passed security or privacy review"]},"engineering":{"fit":"Suits institutions licensing Microsoft 365 Copilot or coding agents with staff already experimenting.","architectureIntegration":"Staff tools reading class schedules and producing attendance PDFs; integration with student and payroll systems unspecified.","prerequisites":["Inventory of staff-built tools","Hosting and authentication standard","Data-classification rules"],"constraints":["Unknown code quality and maintainability","Records-retention rules for attendance and payroll","Different national data-protection law"],"security":"Requires code and access review before tools touch personnel or payroll-related records.","proofOfValue":["Security review of one scaled tool","Baseline versus post-tool processing time","Maintenance hand-off test"]},"delivery":{"work":"Create a pathway from staff-built AI prototypes to supported administrative services.","dependencies":["IT hosting standards","Security review capacity","Unit sponsorship"],"ownership":"Central IT application owner per scaled tool, with HR running the innovation intake.","skillsAdoption":"Pair builder teams with IT developers during hand-off; publish reusable patterns for other units.","governanceCheckpoints":["Data and records review before production","Code review","Annual review of tool ownership"],"acceptanceCriteria":["Proposed: each scaled tool has a named owner and documented data flows","Proposed: measured processing-time change against a pre-tool baseline"],"risks":["Orphaned tools","Unreviewed access to personnel data","Duplicate tools across units"]}},"retrievedAt":"2026-10-13T21:58:37-05:00","enrichedAt":"2026-10-13T22:03:45-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Builds staff AI capability through practice and incentives, run by HR; teams were unit-nominated, so participation may skew toward already-confident staff.","procurementImplications":"Shows build-with-assistants as an alternative to buying point solutions for scheduling, facilities booking and web QA; tool licences (Copilot, Claude Code) become the procurement decision.","operatingModelImplications":"HR-sponsored, unit-nominated innovation with plans to share winning systems across units; scaling requires a hand-off from team-built prototypes to supported services.","sourceVerification":{"openedUrl":"https://www.fcu.edu.tw/en/news/2026100501/","referenceExcerpt":"The new system automatically imports class schedules to avoid conflicts and can generate labor attendance record PDFs that comply with university regulations","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}