Skip to content
Case study · Education

Education
Higher education / academic operations

A college had no shared channel across students, faculty, and staff, and information scattered by role. SDEN deployed University Portal's role-aware AI assistant in five months.

Client
A college
Sector
Higher education / academic operations
Duration
Approximately five months end-to-end

The premise

On most campuses, a student, a professor, and an administrator effectively live in different information universes. There is no shared channel, so a question that should take a message takes a chain of emails, and the tools that could answer it are scattered and role-blind. An assistant that does not understand who is asking is worse than none: it surfaces the wrong things to the wrong people.

University Portal's AI assistant is scoped to each role, with tools that match that role's actual work. This case covers the rollout to a college.

Challenge

No shared channel, information scattered by role

Students, faculty, and staff had no common platform, so coordination ran on email and the answers people needed were scattered across systems none of them fully shared. There was no assistant, and a generic one would have been unsafe, surfacing a professor's tools to a student, or staff data to faculty.

Each role needed different things, and nothing in the existing stack understood the difference.

Approach

An assistant scoped to the role, with the right tools

University Portal deployed five role-tailored experiences, each with an AI assistant wired to role-scoped tools (15 in total across the roles) so the assistant could act only within the asker's permissions, and a shared messaging layer connected everyone.

  1. Phase 1: Role and tool scoping

    Three weeks. Defined what each of the five roles needed the assistant to do, and which of the 15 tools each role could safely use.

  2. Phase 2: Role-aware assistant

    Eight weeks. The assistant deployed per role on University Portal, each instance scoped to that role's tools and permissions, built on Gemini and LangChain behind the platform's access model.

  3. Phase 3: Shared messaging and rollout

    Five weeks. Real-time messaging and notifications connected the roles into one channel, and each role was onboarded to its scoped assistant.

Outcome

A scoped assistant per role, one shared channel

Each of the five roles now has an AI assistant that understands its work and can act only within its permissions, across 15 integrated tools. The chain-of-emails coordination collapsed into one shared messaging channel.

Because the assistant is scoped to the role, it surfaces the right things to the right people, and nothing it should not.

5 roles

each with a scoped AI assistant

15 tools

integrated across the roles

One channel

shared messaging across students, faculty, staff

Let's get to work

Got a project worth building?

Tell us about your project. We work with a limited number of clients at a time, and we'll get back to you within 24 working hours with a first engineer's read, no commitment.

WhatsAppChat with the team
LinkedInFollow SDEN
X@sdenengineering