UFAR · Modeling, Design, Collaborative Development · Group Project

International Student Admission Process

Merge meeting agenda: one shared long-horizon BPM study, from evidence to a proven process, then automation and AI.

30 min meeting Merged group of 5, syllabus max 6 benchmark universities Process first, tech second

Positioning

The line to open with: We picked the process where UFAR loses international students every day. We hold a seed issue register from lived experience and six benchmark universities, verified live. The course grading pipeline maps one to one onto this project. We propose one merged group of five, the syllabus maximum, one shared case study, and a division of labor across the full BPM lifecycle. Process first, proven practices first. Tech and AI second, exactly as the professor asked.

1 Evidence we already hold: seed issue register v0

Every item is first-person observation from living this process as an international student at UFAR. In course terms (Dumas ch. 6.3) this is our starting issue register, to be validated with staff interviews.

#IssueCourse hook
1No English on the website, not even for the English-medium master'sStrategic and customer analysis, PESTEL
2No application system. All files by email. Emails get lostAS-IS: untraceable handoffs, no single channel
3No unified MIS. The website cannot even capture an info requestLecture 1: information system = sender, receiver, channel. Ours is broken
4No callbacks. Phone almost never answered. In-person visit requiredWaste analysis: waiting plus rework. Cycle time
5No reliable or up to date brochureMissing formal IS output. No source of truth
6No confirmation of submissionTraceability, the lecturer's own ISO audit theme
7Armenian or French communication for a 100% English degreeUseless to the customer. Customer relationship failure (A1.2)
8Human factor is the bottleneckCh. 7 bottleneck analysis. Process owner absence
Framing sentence: we are the customer. We hold the complaints of a dozen peers. That is primary data no other group has.

2 Benchmark study: six universities, verified live

Each one shows a different layer of a working admission funnel. None of them wins by technology alone.

GAU

The international marketing funnel
  • Full English prospective section: how to apply, admissions, language requirements, fees, FAQ
  • Application portal apply.gau.edu.tr, applications never touch a personal inbox
  • Brochures in English and Russian, ambassadors, virtual appointment booking, help desk
  • Armenia in the country dropdown: they expect and track applicant origins

WGU

The process discipline model
  • Whole process published as four steps, understood in thirty seconds
  • Named role with SLA: Enrollment Counselor, callable at any point
  • Time-boxed promise: monthly starts, requirements by the 15th to start next month
  • Checklist PDF, free transcript service, direct phone line, request info form

ELTE Budapest

The complete public-university stack
  • Request information form, apply.elte.hu with applicant status visibility, application tutorial
  • Published windows: Nov to Apr to Jul self-funded, Nov to mid-Jan scholarship
  • Ambassadors, in-country tours, local partners, blogs, podcasts, webinars
  • 5,300 international students from 100+ countries at a Hungarian public university

ETH Zurich

Selective-admission transparency
  • Requirement profiles per program with mandatory minimum percentages
  • Committee criteria published in advance: grades, ranking, tests, motivation, references
  • Named Admissions Office, phone, email, published telephone and desk hours
  • Honest limit stated: they will not assess your chances in advance

KTH

The calendar-driven model
  • Six numbered steps with hard dates: open 16 Oct, apply by 15 Jan, docs by 1 Feb, results 1 Apr
  • One national portal, universityadmissions.se, four ranked choices
  • Video guide, program entry requirements, webinars, fairs, campus visits
  • Early submission earns feedback on missing documents before the deadline

TUM

Self-service information plus chatbot
  • Application Info Portal, Enrollment Info Portal, document glossary, FAQ
  • Phone with published hours, service desk walk-in hours, advising by appointment
  • Live student chatbot answering around the clock
  • The chatbot works because the information architecture underneath is clean
What the six prove together
CapabilityProven by
English first, alwaysAll six
Single application channel, never a personal inboxGAU, ELTE, KTH
Process published as numbered steps with datesWGU, KTH, ELTE, ETH
Named process owner with contact and hoursWGU, ETH, TUM
Status visibility for the applicantELTE, KTH
Brochure and checklist maintained and datedGAU, WGU
Self-service info and chatbot on a clean architectureTUM
Proactive recruitment: tours, webinars, partners, ambassadorsELTE, GAU, KTH

UFAR currently has none of the eight rows. That is the gap, and it is a process gap before it is an IT gap.

3 Sequencing: the professor's rule is our design rule

Proven optimized system first, not the tech or IT.

Phase 0
EvidencePESTEL, benchmark study, validation of the issue register with staff interviews.
Phase 1
AS-ISAdmissions and international office interviews. BPMN of the current email and paper flow. Waste analysis, cycle time estimate, Ishikawa.
Phase 2
TO-BE on paperSingle channel, published steps, named owner, time-boxed deadlines, status visibility. Zero technology needed to describe it.
Phase 3
Automation and AI (course-mandated AI use case, layered only on the designed process):
Application portal, single channel, auto confirmation, fixes issues 2 and 6 English web presence, dated owned brochure, fixes issues 1 and 5 Request info form into a tracked queue with SLA, fixes issues 3 and 4 AI voice and chat agent, multilingual, grounded on a temporal knowledge graph of programs, deadlines and fees, escalation to a human with full context DMN eligibility pre-check Process mining once the portal generates logs
The AI sits on a designed process, not instead of one.

4 What we propose to the other group

  • Merge into one group of five (our two plus their three, the syllabus maximum of three to five) on this single case study as a 12 week long-horizon study with real field access.
  • Five owners, one named piece each:
    1 · Strategic and environmental analysis + benchmark study 2 · AS-IS discovery, BPMN, waste and issue analysis 3 · Quantitative analysis: cycle time, bottlenecks, flow analysis 4 · TO-BE, Balanced Scorecard, financial analysis 5 · Automation and AI roadmap, alternative notation, report assembly
    Owners 2 and 3 co-own the analytical phase. Individual defense means everyone explains all parts. Ownership is not a silo.
  • Shared artifacts: one Camunda Modeler repository, one issue register (seed table above is v0), one interview protocol, weekly 30 minute sync, everything timestamped. The lecturer's traceability discipline applied to our own work.
  • Timeline mirrors the course: Ishikawa, process charter and SIPOC, AS-IS, analysis, TO-BE, notations, financials and BSC, AI roadmap.
Ishikawa Charter + SIPOC AS-IS BPMN Analysis TO-BE Notations Financials + BSC AI roadmap

5 Decisions to close in the meeting

  1. Merge, yes or no.
  2. Process boundaries: first inquiry to enrollment decision. Pre-arrival and visa inside or outside scope.
  3. Who books interviews with admissions, international relations, and English program coordinators.
  4. Owner per criterion, list above.
  5. Next session: Ishikawa workshop, bring issue register v0.
  6. Agreement that AI and automation is Phase 3, not the pitch.
Why merge with us: we arrive with a verified issue register and six benchmarks already collected. The topic hits every graded criterion with the mandated AI use case built in. Two groups on the same university means one interview access and two thin studies; merged means one strong study. First person international student evidence is unreplicable by groups studying other processes.