AI-Driven Student Services and Prescriptive Pathways at UCLA Anderson School of Management

ucla · Education · 2024 · source Agentic: yes · promotional bias: med · source-backed · extracted with prompt 0.5.4 (schema v2)

Generated from extraction.json — do not edit by hand. Every claim cites verbatim source evidence; stated means the quote directly asserts the claim, implied means it supports it indirectly.

Problem

Domain: MBA student services and career-path guidance

UCLA Anderson needed to give MBA students personalized, prescriptive guidance across courses, internships, clubs, and career goals while protecting sensitive student data.

Constraints:

Architecture

UCLA Anderson and Kyndryl consolidated disparate student, career, club, and course data into a security-aligned AI application that recommends personalized MBA pathways while keeping institutional policy constraints explicit.

Diagram confidence: reconstructed (reconstructed from prose).

flowchart LR
  student-record-data["Student record data"]
  career-placement-data["Career placement data"]
  club-and-course-catalogs["Club and course catalogs"]
  security-aligned-environment["Security-aligned environment"]
  prescriptive-pathway-application["Prescriptive pathway application"]
  agentic-ai-framework["Agentic AI framework"]
  student-facing-guidance["Student-facing guidance"]
  student-record-data --> prescriptive-pathway-application
  career-placement-data --> prescriptive-pathway-application
  club-and-course-catalogs --> prescriptive-pathway-application
  security-aligned-environment --> prescriptive-pathway-application
  agentic-ai-framework --> prescriptive-pathway-application
  prescriptive-pathway-application --> student-facing-guidance
Drag to pan · scroll or use +/− to zoom · reconstructed from the source's described component topology

Components

Component Role Interacts with
Student record data Provides academic and profile context for personalized pathways. Prescriptive pathway application
Career placement data Connects student goals to employers, internships, and outcomes. Prescriptive pathway application
Club and course catalogs Supply recommendations for activities and coursework. Prescriptive pathway application
Security-aligned environment Mirrors UC information security policy for student-data handling. Prescriptive pathway application
Prescriptive pathway application Synthesizes records, career data, clubs, and courses into recommendations. Student-facing guidance
Agentic AI framework Supports future agent-to-agent handoffs and workflow orchestration. Prescriptive pathway application
Student-facing guidance Presents courses, companies, internships, and club recommendations. Prescriptive pathway application

Data flow

The pathway system pulls from student records, career placement, clubs, and course catalog data before generating career-aligned recommendations.

“we took from a lot of different disparate systems. We took from the student's record database, we took from our career placement, we took from our clubs, we took from our course catalog.”transcript, stated

Why this architecture

The environment was designed around UC information-security policy because sensitive student data made this more than a casual AI pilot.

“we took extra care to architect that environment such that it mirrored the Information Security Policy of the UC system almost line for line so if it was ever gonna get audited that it would pass that test.”transcript, stated
“we wanted to ensure the security and privacy of that student data. So we took extra care to architect that environment such that it mirrored the Information Security Policy”transcript, stated

Models used: Kyndryl agentic AI framework

Scale

Alternatives considered (roads not taken)

Shipping quickly before architecture and policy review

Rejected because: The team spent substantial time getting the architecture right, even though the speaker later wished they had failed faster.

“I think I would have failed faster to begin with. I think we spent so much time trying to get the architecture right that by the time we launched, the technology had changed so much”transcript, stated

Point solution over one campus dataset

Rejected because: The system integrates student records, career placement, clubs, and course catalogs instead of relying on one data source.

“we took from a lot of different disparate systems. We took from the student's record database, we took from our career placement, we took from our clubs, we took from our course catalog.”transcript, stated

Security review after product build

Rejected because: The environment mirrored UC information-security policy so student-data controls would pass audit.

“we took extra care to architect that environment such that it mirrored the Information Security Policy of the UC system almost line for line so if it was ever gonna get audited that it would pass that test.”transcript, stated

Patterns

rag-retrieval (student-services)

The application combines student records, career placement, clubs, and course catalog data to recommend a personalized roadmap.

“we took from a lot of different disparate systems. We took from the student's record database, we took from our career placement, we took from our clubs, we took from our course catalog.”transcript, stated

safety-policy-enforcement (governance)

The environment was architected to mirror UC information-security policy so sensitive student-data use could pass review.

“we took extra care to architect that environment such that it mirrored the Information Security Policy of the UC system almost line for line so if it was ever gonna get audited that it would pass that test.”transcript, stated

Tradeoffs

How quickly to ship a student-data AI application.

Chose security-first architecture and consolidation over faster release with less governed data.

Rationale: The team later wished they had failed faster, but the chosen path reflected the sensitivity of student records and campus audit requirements.

“I think I would have failed faster to begin with. I think we spent so much time trying to get the architecture right that by the time we launched, the technology had changed so much”transcript, stated

Gotchas (what bit them)

Sensitive student data made the project harder than a quick AI pilot.

Root cause: The application required third-party risk management, privacy review, and data consolidation across multiple systems.

Resolution: Mirror UC security policy and build the application around governed student-data access.

“we wanted to ensure the security and privacy of that student data. So we took extra care to architect that environment such that it mirrored the Information Security Policy”transcript, stated

Evaluation

The case reports delivery timeline, production-use-case progression, and student-pathway utility rather than a formal accuracy benchmark.

Metrics: 2 million paper records digitized · first two production use cases · approximately eight-month project timeline

“at UCLA, which we'll dig into deeper during our fireside chat, we have worked on digitizing over 2 million paper records and laying the foundation for AI-driven student services.”transcript, stated

Outcomes

Education/government conference case; outcomes are qualitative and client-reported.

“If this is what you want to do, you should take these courses. You should look at these companies for internships. You should join these clubs.”transcript, stated