This introduction summarizes the fast take: A Jan 2025 design-thinking workshop at UNC produced stakeholder-driven ideas on GenAI, remote interviews, and admissions criteria, with n=15 participants generating about 76 ideas and presenting 4 prototypes; read the original study at PLOS ONE. This post shows how to convert those generative outputs into defensible qualitative evidence with an AI-enabled workflow and links to Evidano for artifact ingestion, reproducible thematic analysis, and shareable visualizations for decision-makers.
Key Takeaways
Evidano is an AI-powered qualitative data analysis platform that converts design-thinking workshop artifacts into auditable, reproducible thematic syntheses to inform admissions pilots and policy.
Design thinking workshops can rapidly surface many actionable ideas, and an AI-enabled synthesis turns those artifacts into defensible evidence for decision-makers.
- A Jan 2025 UNC DT session (n=15) produced ~76 brainstormed ideas, 43 ideas prioritized across 3 themes, and 4 prototypes, demonstrating high generative yield.
- The UNC study used inductive thematic coding with 90.5% inter-coder agreement and a 94% post-workshop survey response rate, useful benchmarks for synthesis quality.
- Teams can ingest photos, facilitator notes, and vote tallies to produce cross-segment comparisons and audit trails before piloting changes.
Findings snapshot
| Date | Event | Sample | Raw ideas | Selected | Prototypes | Source/Note |
|---|---|---|---|---|---|---|
| January 2025 | 2‑hour DT workshop (UNC Eshelman School of Pharmacy) | n = 15 (faculty, staff, reviewers, student ambassadors) | ~76 brainstormed ideas | 43 ideas prioritized across 3 themes | 4 prototypes presented | PLOS ONE |
What happened: the study in plain terms
The study ran a structured design-thinking (DT) session with purposively sampled admissions stakeholders to reframe problems, generate divergent solutions, and draft low-fidelity prototypes.
The study collected physical artifacts (chart paper, worksheets), facilitator notes, and a brief post-workshop survey that achieved a 94% response rate.
Analysis in the study used inductive thematic coding with 90.5% inter-coder agreement, while quantitative responses were reported descriptively (for example, creative problem solving M=4.7 ±0.5).
- Primary focal areas were interview restructuring, GenAI integration/compliance, and broadening admissions criteria.
- Examples of ideated tactics included task-based live writing stations to preclude GenAI, technology-mediated monitoring, GenAI-assisted rubric generation, and adding timed writing to candidate days.
- The study was generative and exploratory, single-site with a small purposive sample, appropriate for ideation rather than outcome claims.
So what for admissions researchers and UX teams
Admissions directors
Admissions directors should use DT to surface locally relevant trade-offs such as cost, geography, and equity before committing to pilots.
The UNC workshop shows how focused, two-hour sessions can produce many actionable leads quickly.
Admissions directors should prioritize ideas that map to measurable outcomes, for example reduction in suspected GenAI use or equity metrics for interview completion.
UX / research teams
UX and research teams should treat DT artifacts as qualitative data: transcribe, code, and compare themes across stakeholder roles to reveal alignment or friction.
Teams should capture vote counts and prototype rationales (the study reported vote ranges 0–13) as metadata to weigh stakeholder buy-in during synthesis.
Policy & ethics leads
Policy and ethics leads should embed legal and accessibility constraints such as ADA in the DT constraints, as UNC did with cost, geography, and equity.
Policy and ethics leads should design for auditability by preserving raw artifacts and coded outputs to document why a process change was selected or rejected.
Do more, faster with Evidano (map to this use case)
From artifacts to reproducible themes
Evidano enables mapping workshop artifacts to reproducible themes by ingesting photos of chart paper, worksheets, and facilitator notes and auto-transcribing typed and handwritten text to searchable text.
Evidano can run inductive thematic synthesis and export a hierarchical codebook (themes → subthemes) with coder agreement metrics, helping teams replicate the study’s 90.5% inter-coder check faster.
Compare segments and weigh votes
Evidano lets teams import vote counts and participant roles to produce cross-segment frequency tables so teams can see which prototypes had broad support.
Evidano’s cross-segment analysis helps identify aligned versus contested ideas before piloting, making stakeholder trade-offs visible.
Prototype validation without the travel budget
Evidano supports remote prototype validation using AI avatar interviewers to simulate candidate responses and structured critique tasks such as evaluating a GenAI-generated essay.
Evidano captures structured feedback and feeds it back into the platform for iterative coding and visualization.
Secure, auditable, research-grade
Evidano provides encrypted customer data handling and proprietary LLMs tuned for qualitative research, and customer data is never used to train third-party models, suitable for sensitive admissions material.
Evidano exports reproducible reports including word clouds, co-occurrence networks, and hierarchical code maps for committee briefings and IRB documentation.
Checklist: 7-step workflow to reproduce the UNC DT synthesis in Evidano
This checklist provides a seven-step runbook to convert a DT session into decision-ready insight.
- 1) Collect artifacts: photos, worksheets, facilitator notes, and vote tallies, then export them into one folder.
- 2) Ingest files into Evidano and run auto-transcription with an optional custom dictionary for domain terms.
- 3) Auto-suggest initial codes and import your manual codebook; run AI-assisted coding to tag quotes and artifacts.
- 4) Compute code frequencies, cross-segment comparisons (role, vote count), and co-occurrence networks.
- 5) Validate top themes with member-checking by generating a one-page summary and running a short AI-facilitated follow-up with selected stakeholders.
- 6) Prototype remotely by using AI avatars to simulate candidate tasks (for example, critique a GenAI essay) and collect structured responses.
- 7) Produce an exportable synthesis (slides plus an appendix of raw quotes and an audit trail) for the admissions committee.
FAQ: design thinking in admissions and AI-enabled qualitative analysis
When should I use design thinking versus a survey or formal experiment?
Use design thinking for early, generative phases to reframe problems and produce prototype concepts.
Design thinking is ideal when teams need to surface a wide range of ideas quickly; use surveys or experiments later to evaluate effectiveness and outcomes.
How do I compare stakeholder votes fairly?
Compare stakeholder votes by recording role metadata and vote counts as structured fields.
Evidano’s cross-segment analysis highlights where support concentrates and where perspectives diverge, enabling fairer comparisons.
Is it ethical to analyze admissions artifacts with AI?
Analyzing admissions artifacts with AI can be ethical for research and process design if privacy standards are followed.
The UNC study was deemed not human-subjects research; for human data involving admissions decisions, consult your IRB and follow consent and data-handling protocols.
Wrapping up: next moves for admissions teams
Design thinking produces rich, stakeholder-driven ideas quickly, and the UNC workshop published June 23, 2026, shows how a small, well-structured session can yield dozens of actionable leads (n=15, ~76 ideas).
- Run a DT session, ingest artifacts into Evidano, and produce an auditable thematic synthesis plus cross-segment comparisons.
- Pilot one low-cost prototype (for example, a timed writing MMI station) and measure candidate experience and equity outcomes before scaling.
- If you want to try this pipeline, Try Evidano for free or learn more at Evidano and import your first workshop to generate a shareable synthesis in hours, not weeks.