From the Campus Operations edition of November 3, 2026
Pusan National University trains 40 administrative section managers to build AI agents after staff built 12 work agents in an eight-week course
Pusan National University (AX and Informatization Innovation Headquarters) · Higher education administrative operations · Busan, South Korea
- Publisher
- Busan Ilbo (busan.com)
- Original publication
- 3 November 2026 (Busan Ilbo)
- Source retrieved
- 2026-11-04
- Event date
- 2026-11-03
What happened
On 2-3 November 2026 Pusan National University's AX and Informatization Innovation Headquarters ran a six-hour course for about 40 administrative section and office managers: day one on how generative AI works in administrative tasks, day two building work AI agents with agentic tools, plus the amended public-sector AI law in force since 28 August. It followed the 'AI Maestro' staff course (19 June-15 October), in which 21 staff from 20 departments built 12 work agents over eight weeks; self-assessed AI skill rose from 2.85 to 3.94 out of 5.
Why it matters
Shows a campus extending agent-building from front-line staff to middle managers so they can judge which departmental work to give to AI.
Evidence and measured results
News report of a university release; outcome data is self-assessment only and the 12 agents are not described or measured.
Limitations and uncertainty
Self-reported skill gains only; no productivity, quality or error data; agent inventory not published.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-11-03; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Departments build AI agents but managers lack criteria to judge which tasks to delegate.
- Stakeholders
- AX/IT office head
- Section managers
- Department staff
- Discovery questions
- How many staff-built agents exist and who approves them?
- What review criteria do managers use?
- How is compliance with AI law checked?
- Value hypothesis
- Manager training may improve selection of agent use cases; no measured benefit is reported.
- Potential engagement
- A bounded agent review framework and inventory for departmental agents.
- Unsupported claims
- Productivity or time savings
- Quality of the 12 agents
- Durable skill gains beyond self-assessment
Pre-sales engineering
Role takeaway
- Fit
- Campuses with citizen-built agents in administrative departments.
- Architecture and integration
- Agentic tool platform not named; departmental data sources not described.
- Prerequisites
- Approved agent-building platform
- Data access rules
- Agent registry
- Constraints
- Korean public-sector AI law obligations
- Unknown hosting/jurisdiction of tools
- Security
- Need access scoping and logging for staff-built agents; none described.
- Proof of value
- Inventory and classify the 12 agents by data used
- Measure task time before and after for a sample agent
Delivery
Role takeaway
- Work
- Run role-tiered agent training and move from training outputs to governed departmental agents.
- Dependencies
- Central AX office
- Department managers
- Legal/compliance
- Ownership
- AX and Informatization Innovation Headquarters.
- Skills and adoption
- Hands-on build labs for staff; evaluation-focused labs for managers.
- Governance checkpoints
- Manager review of staff-proposed agents
- Legal check against public-sector AI law
- Acceptance criteria
- Proposed: every production agent registered with owner and data scope
- Proposed: outcome measure defined per agent
- Risks
- Unmanaged agent sprawl
- Overreliance on self-reported gains
- Data exposure through unvetted tools
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Staff-built agents on agentic tools imply many small departmental agents; platform and hosting are not named.
Governance
Who approves, reviews and stays accountable for outcomes?
Managers are trained to review staff-proposed AI uses and to check obligations under the amended public-sector AI law.
Security and privacy
What data, permissions and controls need testing?
No data-handling controls for the staff-built agents are described.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Role-tiered training: staff build, managers assess fit; satisfaction survey planned.
Procurement
What should contracts, pricing and exit terms secure?
Tooling vendor not named; scale-up would raise licensing and support questions.
Operating model
Which teams own the service once it runs?
Central AX office runs training; departments own use cases.
Publication history
- 2026-11-03Campus Operations · Issue 563 resources
Stable resource ID: pusan-national-university-manager-ai-agent-training-20261103