← All case studies

Applied AI / 3 min read

EMUHelper

Give students a useful next step, even when the answer is unclear.

A student support prototype exploring common questions, source-based guidance, and a clear handoff to human support.

Context
Independent university support prototype
Audience
Students and support staff
My focus
Support journey and escalation

01 / Context

The problem and my role

Students may need help with registration, fees, documents, or university services without knowing which office or source to consult. The product concept brings common topics into one starting point.

I worked on the problem definition, support journey, knowledge-base structure, and prototype. The experience demonstrates a conversation, quick topics, and a ticket summary for cases that need more help. It is independent portfolio work.

Student

Find relevant guidance and understand what to do next.

Support staff

Receive a clear account of the issue instead of restarting the conversation.

The experience at a glance

  1. 01Choose a topic or ask a question
  2. 02Find relevant guidance
  3. 03Clarify an uncertain request
  4. 04Prepare a handoff when needed

02 / Product judgment

The choices that shaped the work

The central question

When should an assistant answer, clarify, or hand the question to a person?

01Make frequent topics easy to start

Offer quick actions for common student support needs.

Why it matters
A student should not need to invent the right prompt before receiving help.
The tradeoff
Topic coverage and wording need testing against the questions students actually ask.

02Treat uncertainty as part of the journey

Ask for clarification or escalate when the available information is not enough.

Why it matters
A plausible answer can still send a student to the wrong place.
The tradeoff
Escalation adds work for staff, so both answer quality and handoff quality need evaluation.

03Preserve the issue in a summary

Collect the relevant issue details and demonstrate a ticket summary.

Why it matters
The next person needs context to continue helping.
The tradeoff
A generated summary does not establish that a university office received or resolved a ticket.

03 / Evidence

What the prototype makes testable

The product concept separates a browser demonstration from the service operations a university would need.

JourneyDemonstrated behaviorOperational requirement
Common questionsQuick topics and a conversational support flowApproved, maintained sources and answer-quality review
Unclear requestsClarification and escalation pathRules for what staff should receive and how quickly
HandoffIssue and ticket summaryA connected service desk, ownership, and confirmation of receipt

04 / Evaluation

What I would measure and why

Measures to review or validate next. These describe the evaluation approach, not achieved improvements.

MeasureWhat to examineWhy it matters
Answer qualityReview answers against approved university information.Check correctness and relevance before optimizing for fewer escalations.
Task resolutionAsk whether the student found the correct next step.A completed conversation is not necessarily a resolved issue.
Handoff usefulnessHave support staff review sample issue summaries.Check whether the summary contains enough context to act.

05 / Outcome

Where the work stands

The result is an accessible prototype for reviewing the support experience, with quick topics, conversation flow, and escalation summaries. Live university integration and measured service improvements are outside the demonstrated outcome.

What comes next

  1. Evaluate a set of common, ambiguous, outdated, and out-of-scope questions.
  2. Assign ownership for source updates and define when an answer must be reviewed.
  3. Validate ticket routing with staff before presenting the handoff as an operational university service.

My takeaway

A support product needs a good path through uncertainty. Knowing when to clarify or hand over is part of providing a useful answer.

Based on the student support concept, prototype scope, and portfolio documentation.

Continue exploring

Applied AI Systems

A showcase of how I structure prompts and workflows around a clear user task.

Next case study →All projects