in Business Intelligence

Designing Authentic Industry-Engaged Assessment for Professional Competence

Dr. Helga Ingimundardóttir helgaingim@hi.is

22^nd International CDIO Conference · June 23^rd, 2026

I’m Helga Ingimundardóttir from the University of Iceland — this is my first time at CDIO, and I’m presenting twice this week.

• Today: why we redesigned and what the design looks like
• Tomorrow (June 24): how — pull requests as assessment infrastructure
• Relevant if you do industry-engaged or project-based courses

Motivation

Academia

aims to prepare students for professional practice, but offers limited guidance on how to operationalize this in course design and assessment

Industry

expects credible, reproducible results supported by clear reasoning and usable for decision-making

Underlying mismatch

what counts as “good work” differs between academic and professional contexts

Everyone agrees academia should prepare students for professional practice — the question is how to actually design for it.

• This isn’t a values mismatch — it’s a design mismatch
• Industry needs work that is credible, reproducible, decision-usable
• Our own course: the partner valued the collaboration but would not use the results; students found it heavy and uneven
• Core question: how do we design so work is authentic in practice and meaningful for learning?

Context: Business Intelligence

Overview

• 6 ECTS course
• advanced undergraduate and early graduate students
• 10–15 students, two teams

What is BI?

data-driven decision support

• data analysis
• systems integration
• communication to inform organizational decisions

Goal: strengthen industrial relevance and professional competence

Business Intelligence is about decision support — not just producing correct results, but communicating assumptions, limitations, and implications.

• 6 ECTS, advanced undergrad and early grad, 10–15 students in two teams
• Small enough for sustained industry engagement — that’s the key

Industry-engaged project

Sustained industrial context

a single dataset and domain used across the entire semester

Industry partner role

• problem framing
• milestone discussions
• final evaluation

Dual context

academic and industry perspectives coexist throughout

The biggest single structural change was using one dataset, one domain, all semester.

• No repeated onboarding — students build depth instead of breadth
• Industry partner present at problem framing → milestones → final evaluation
• That continuity is what creates conditions for usable results

Design focus: LOUIS-guided

Focus on a small set of competences within a sustained industry context.

LOUIS — Learning Outcomes in University for Impact on Society (Aurora University Network). Focus on a small set of competences (max 3).

To focus the redesign, I used the LOUIS framework — Learning Outcomes in University for Impact on Society — developed by the Aurora University Network, a European alliance of nine research universities of which the University of Iceland is a member. Aurora runs faculty workshops to help teachers embed these competences into their courses — which is exactly what I did.

• Key workshop insight: don’t spread thin — pick at most three competences and make expectations explicit
• We chose: Information literacy · Problem solving · Teamwork

Pedagogical structure

Team-Based Learning: iRAT (individual readiness)  ·  tRAT (team readiness)  ·  tAPP (team application)  ·  iREF (individual reflection)

Flipped classroom  ·  Iterative modules leading into capstone  ·  iRAT/tRAT use is slightly unconventional

Every module follows the same rhythm: prepare → align → apply → reflect — this is exactly what students will meet in the workplace as the Scrum process: sprint planning, sprint, retrospective. The course structure itself is professional practice.

• iRAT / tRAT: readiness checks — everyone arrives prepared
• tAPP: team applies to the shared project — where the real work happens
• iREF: individual reflection — what changed, what you’d do differently
• After all modules, a mini-capstone bridges to the final capstone — which is not a new project but a revisit and refinement of everything that came before

Assessment as development

Core mechanism

One rubric, many iterations

same assessment rubric reused across all modules and the capstone

• analytical soundness
• transparency
• reproducibility
• stakeholder relevance

Developmental shift

feedback is carried forward — not reset between tasks

Standards fixed; performance expectations scaled during the learning phase.

Students see the final rubric from day one — the standard doesn’t change, but expectations are scaled during the learning phase.

• In practice this means pull requests: changes proposed, reviewed, integrated — or explicitly deferred
• Unresolved issues stay visible across submissions — nothing disappears
• Same rubric reused → grading is faster; I can see exactly what changed

Evolving feedback

Instructor

• method
• implementation

Industry partner

• domain relevance
• decision context

Student

• knowledge transfer through PRs
• defending decisions
• critiquing peer work

Progression over time

• align within teams
• cross-team review
• industry validation

Audience shift

from internal review to external scrutiny

Feedback comes from three sources, each looking at something different — and the audience widens deliberately over the semester.

• Instructor → method and implementation
• Industry → domain relevance and decision context
• Students → defending their own decisions, critiquing peers
• Scrutiny increases: within team → cross-team → industry validation
• Full technical implementation is tomorrow’s talk

Credibility and trust

Industry threshold

credibility is fragile: early issues can undermine trust in the entire result

What is evaluated

• reproducibility
• transparency
• clear assumptions

Shift in communication

build trust first — then invite inspection

Students are told to present to sell — then invite the partner to read the codebase if curious, and the final report if they want the full picture. (The report is what I grade — the partner never needs to.)

Let me give you a concrete example of why credibility matters more than you might expect.

• Our industry partner found one error early in a student presentation
• The rest of the work was solid — but they lost trust in all of it and stopped engaging
• That experience drove the shift: build trust first, then invite scrutiny
• Reproducible workflows are how you make credibility visible — not just claimed

Indicative evidence

Students & graduates

• high engagement without attendance requirements
• transfer of reproducible workflows
• stronger documentation practices

Academic perspective

more ambitious, student-driven projects

Industry perspective

improved coherence and credibility; results could be implemented in practice

This has only been run once — so take it as a design case, not a controlled study. But the early signal is promising across all three groups.

• Students: no attendance requirement, yet highest attendance of any course I teach this cohort
• Graduates: they transfer the workflows — reproducible practices and structured documentation, years later
• Industry: outputs are more coherent and usable — results that could actually be implemented
• Personally — this is the most enjoyable course I teach. I’m just a facilitator; the students do the work. They’re also the cohort most likely to knock on my door for advice — that says something.

What this design makes visible

Transferable principles for CDIO-aligned, industry-engaged courses.

Sustained context

one domain, all semester — depth over breadth

Assessment as iteration

same rubric, feedback carried forward — not reset

Visible contributions

accountability without fragmenting the team

Widening audience

from internal review to industry scrutiny

Accumulated feedback

unresolved issues stay visible across submissions

Tooling as support

design comes first — tools enable it

These six principles apply beyond Business Intelligence — the domain matters less than the design.

• Students often find this course uncomfortable in the moment — but graduates tell me it’s where they learned the most, because of the hands-on experience
• They see it when applying for jobs: the skills are real and they know it
• Critical feedback, visible work, structured workflow: better to meet them first in a course than on the job

Questions?

Dr. Helga Ingimundardóttir

helgaingim@hi.is

0000-0002-2780-3546

tungufoss

University of Iceland

I’m happy to take questions — and if limitations come up: yes, it depends on student readiness, it scales to around 15 students, and domain transferability is still open.

• Active refinement: strengthen the initial phase — more time to build trust before methods get complex
• Tomorrow, June 24 at 1pm — Session 8B: Learning Environments, Labs and Simulations — the technical side: “Using Pull Requests to Make Collaboration Visible”

Team and individual accountability

Team-level

• analytical quality
• coherence
• relevance

Individual-level

• visible artifacts
• documented contributions
• structured reflection

Design goal: accountability without fragmenting teamwork — designed into the course architecture, not added after the fact.

The classic tension is teamwork versus individual accountability — and the usual fix of splitting tasks into isolated pieces breaks collaboration.

• Our approach: keep the team together but make contributions visible
• PRs show who did what, how they engaged with feedback, how their thinking evolved
• Accountability designed in — not bolted on after the fact
• Full detail in tomorrow’s talk