Faculty ProjectsImproving the Student Advising Experience with an AI Advising Agent
Faculty
Chamille Lescott, BME
Amount Requested
$1,525
Summary
In McCormick, tenure-track faculty are responsible for undergraduate advising in most departments. As students increasingly combine the BME curriculum with minors, certificates, pre-med preparation, study abroad, and other academic opportunities, the number of feasible pathways through the curriculum has expanded dramatically. Reviewing a student's four-year degree plan has become increasingly complex, requiring advisors to evaluate prerequisite chains, course sequencing, course offerings, and degree requirements across many possible paths.
Although McCormick's current degree audit system effectively tracks progress toward completed degree requirements, it does not evaluate whether a student's proposed multi-year plan is feasible. As a result, students may unknowingly create schedules that violate prerequisite requirements, rely on courses not offered in a particular quarter, or postpone critical courses until it is too late to remain on track for graduation. As a result, faculty and students spend much of their advising meetings verifying degree plans rather than discussing students' academic interests and career goals.
This creates an opportunity to reimagine the advising experience by leveraging AI.
The Biomedical Engineering Advising Partner (BMAP) is an AI agent that reviews students' proposed four-year degree plans, identifies potential scheduling issues before they become problems, and prepares both students and faculty for more productive advising meetings.
Specifically, BMAP:
- Provides students with useful information about the errors they've made with their degree plan, which may include selecting courses not offered in that term, miscounting credits towards a requirement, or ignoring prerequisite information.
- Directs faculty towards the most important issues with a degree plan, reducing the time spent verifying the study plan.
- Shifts advising conversation away from course selection and course sequencing and towards deeper conversations about students' goals and interests.
- Reduces administrative work by identifying planning problems early, decreasing the need for later schedule revisions, exceptions, and graduation workarounds.
We envision students using BMAP before meeting with their faculty advisor, iteratively refining their degree plans and identifying topics for discussion. Rather than replacing faculty advising, BMAP augments it by handling routine curriculum analysis so faculty can devote more time to mentoring students.
This project serves as a pilot for a new advising tool in Biomedical Engineering. If successful, it can be adapted to other McCormick departments, improving advising efficiency while enabling faculty to focus on higher-value mentoring.
Planned Activities/Investments
BMAP will be launched in 3 stages. We have currently created a functional prototype of BMAP in MindStudio, a no-code platform for building and launching custom AI agents. The tool is powered by Claude. We are careful to avoid asking students for any FERPA-protected information.
The current agent takes roughly 2 minutes to run and can analyze 4-year degree plans from multiple catalog years. The cost per run of the agent varies based on the complexity of the student's degree plan, but the average cost is between $0.60 and $0.80 per run.
Project Development Stages:
Stage 0: Agent refinement (completed Spring 2026): Tested AI Agent on student degree plans collected over the 2025-2026 academic year.
Stage 1: Small group testing (Summer 2026): Collect user feedback from rising junior and senior students. Make any necessary changes to the BMAP agent architecture.
Stage 2a: Launch BMAP with first-year students in the BMD_ENG 101 course, a first-year seminar that introduces students to BME. (Fall 2026, Winter 2027)
Stage 2b: Launch BMAP for all sophomore, junior, and senior students in BME. Collect feedback at each stage and continue refining the AI agent logic. (Fall 2026, Winter 2027, Spring 2027)
Stage 3: Full launch of BMAP for all BME students (Spring 2027)
Impact
This project will impact every undergraduate in BME, the faculty who advise them, and the advising relationship itself:
- Students. All undergraduate BME students (currently about 180 to 200 students) will gain on-demand access to accurate, individualized feedback on their own degree progress, instead of depending on the few faculty members who can reliably track every requirement.
- Faculty. Every faculty advisor, not just the small group who currently absorb most of the advising load, can advise with confidence, because BMAP surfaces the specific issues in a student's plan instead of expecting each advisor to track every curriculum detail from memory.
- Student-faculty relationships. As advising time shifts away from course logistics, students and faculty gain room for "life planning" conversations, in which faculty learn about student goals and provide guidance that builds a lasting mentoring relationship rather than a transactional one.
We will evaluate this impact at each stage through interviews, surveys, and focus groups, tracking whether students report more confidence navigating degree requirements, whether faculty report more confidence advising, and whether advising conversations shift toward goals and life planning rather than course logistics. Details by stage are below in the Evaluation Plan.
Evaluation Plan
Stage 1: Small group testing with 10 students who are following the 2025-2026 requirements; Summer 2026 – We will interview each student about their experience with BMAP and collect feedback. At this stage, we'd also like to know how students think BMAP will impact sessions with their advisor. At each evaluation stage, we can view BMAP analytics, including cost and time per run, and track common errors in student plans. We can encourage students to rigorously test the tool by running the AI agent multiple times, so they can provide quality feedback.
Stage 2a: BMD_ENG 101 testing (2026-2027 requirements); Fall 2026, Winter 2027 – We will collect study plans from students enrolled in the BMD_ENG 101 seminar, coach the students on how to use BMAP, and survey them at the end of the experience. At this stage, we want to know how BMAP engagement helped them understand BME requirements and whether there are any pain points associated with the BMAP process.
Stage 2b: Testing with current BME students (2025-2026 requirements), Fall 2026, Winter 2027, Spring 2027. At the end of each quarter, we will send out a survey to faculty and students to collect feedback. Faculty reactions to BMAP will be collected during full-group faculty meetings. Here we can probe the 3 core goals of promoting (1) student awareness of requirements, (2) faculty confidence with advising, and (3) opportunities for student-faculty connections.
Stage 3: Full launch of BMAP for the whole BME student population (Spring 2027).
Deliverables
The deliverable is a working AI advising agent for Biomedical Engineering, plus a development strategy and architecture that other McCormick departments can follow to build similar tools.
Sustainability
Once we've demonstrated the usefulness of this tool, we will be able to incorporate the costs of sustaining BMAP into the BME annual budget. BME Business Administrator, Deb Cundiff, confirmed that the relatively low annual cost of BMAP (<$2000) was a feasible expense to budget for once the tool was deemed a departmental priority.
Budget Overview
MindStudio charges $20/month to host the agent. However, at this $20 /month cost, we can develop and launch multiple agents. ($240/year)
Right now, the BMAP degree plan analyzer agent costs ~ $.70/run (depending on the complexity of the student’s study plan)
To cover the developmental costs of stages 1 and 2, we estimate the following token costs:
To conservatively estimate token costs and plan for the possible increase in complexity of the BMAP tool as the agent is refined, we will assume each run costs ~$1.00.
- Stage 1: 100 runs, conservatively estimate $100 in token costs as students vet the tool
- Stage 2a: 225 runs, conservative estimate $225 in token costs, as ~75 students repeatedly use the tool in BME 101 (~3 runs per student).
- Stage 2b and 3: Assuming we have 200 students who run the tool at least 3 times per year (quarterly), and we acknowledge that some may run the tool multiple times per quarter (2-3 runs per student), this leads to approximately $1200 in token costs per year when we launch Stages 2 and 3 of the project.
Total Budget Amount: $1,525
Matching Funds
The BME department is prepared to cover the cost of hosting the AI agent ($20/month, $240/year) to demonstrate the department's support and commitment to this project.