There's a version of the AI in higher education conversation that stays at the level of principle. Governance matters. AI should be embedded, not bolted on. Institutions need to maintain control over content and data.

All of that is true. But it stays abstract until you see what it looks like when an institution actually does it.

Below, we dive into three recent partnerships we’ve begun with different institutions, across platforms and various use cases that show how embedding AI solutions like LearnWise works in practice.

Pima Community College: Support that lives where learning happens

Pima Community College serves over 40,000 students across Southern Arizona, including a significant number of fully online learners who log in to Brightspace late at night, when institutional staff aren't available. Their challenge wasn't a shortage of information: it was accessibility in context. Students and faculty often couldn't find what they needed quickly, or couldn't articulate the problem well enough to navigate fragmented support systems.

In early 2026, PCC deployed Lumi Chat (powered by LearnWise) directly inside Brightspace, alongside a planned integration with TeamDynamix for structured ticket intake. The goal was to create support that didn't require leaving the learning environment: a student encountering friction inside Brightspace could ask a question in plain language, receive an immediate answer when possible, or have a ticket created on their behalf without needing to understand which system was responsible for which problem.

As Tony Sovak, LMS & eLearning Quality Director, described it: if users are going to turn to AI for help, it should happen inside a system aligned with the institution's values, governance, and support structures. Here, it is not about preference, but about making conscious governance decisions.

Read the Pima Community College story

American College of Financial Services: Program-level AI tutoring for working professionals

The American College of Financial Services serves 17,000–18,000 students annually, all of them working professionals studying for financial designations entirely online. Their learning happens around full-time careers, which means most questions arrive outside office hours, and advising teams were spending significant capacity on repetitive inquiries about exam registration, course navigation, and program requirements.

In late 2025, ACFS deployed SPARK, their branded instance of LearnWise's AI Student Tutor, inside Brightspace. What distinguished the implementation was its scope: rather than building a chatbot for individual courses, SPARK was trained across entire designation programs, so it could support students across their full credential journey rather than in isolated modules.

The early results were measurable. In the first 60 days across three courses, students asked over 700 questions. Of the more than 900 responses SPARK delivered, only two required escalation to a human. Faculty who tested SPARK with complex financial planning scenarios - questions that had previously taken hours of manual research - found it delivered structured, accurate responses in minutes.

One student summed up the shift simply: SPARK became their go-to over ChatGPT. That's what it looks like when AI is trained on institutional content rather than the open web.

Read the American College of Financial Services story

University of Westminster: Embedding AI inside Blackboard at institutional scale

The University of Westminster has used Blackboard as its virtual learning environment for over 15 years. With 21,000+ students across four London campuses, their goal was not to experiment with AI, but to deploy it in ways that were safe, equitable, and useful, aligned with a clear institutional policy they had been developing for two years.

They deployed LearnWise inside Blackboard and in their Student Centre, providing students with an AI assistant available at any time, trained on the content of each specific Blackboard course site. Since the start of the 2025–26 academic year, approximately 34,000 queries were submitted to the Student Centre, with 9% handled by the AI assistant. Every single audited response was accurate.

What emerged alongside the support improvements was something less expected: early-stage interest from the law school in using anonymous interaction data to understand where students struggle and how they approach topics. AI embedded in the LMS isn't just reducing support load. It's starting to reveal things about how learning works that weren't visible before.

Read the University of Westminster story (via Blackboard)

What these three institutions share

These institutions are different in scale, platform, student profile, and geography. But they share a few consistent decisions: they started with a specific, operational problem rather than a general ambition; they kept AI inside the systems students and staff already use; and they built governance and content control into the deployment from the start, rather than adding it after the fact.

The bottom line: Responsible AI adoption is about making institutional decisions about where and how AI should belong, and how to make sure it is well-managed. This is the best way to scale and expand ethical AI use beyond one proof of concept or a one-off pilot sitting in one department.