From Pilot to Practice: Building an Institution-Wide AI Support Strategy at IE University

Overview

IE University's IT Innovation team evaluated several AI assistants before finding one that could scale past a single use case. After early deployments stalled on platform and scaling constraints, the team adopted a federated model built on LearnWise AI's AI Campus Support, handing agent ownership to individual program teams rather than managing it centrally. Since December 2024, IE's portfolio of these assistants has handled 3,754 conversations and 3,886 student and staff messages, with a 94% resolution rate on average.

Context & Challenge 

IE University's IT Innovation team needed a way to deliver support across many different programs, intakes, and audiences without the burden falling on one small central team, or creating the massive project of centralizing all knowledge from many different teams into one cohesive knowledge base. The team had tried other AI assistants before LearnWise, and each one ran into the same pattern. In Milena Janowska's words, EdTech Innovation Manager at IE University: "Every time we had some issues with integration with our systems, or it was difficult to implement, or it was very difficult to train. It was taking ages."

AI Campus Support is LearnWise's student- and staff-facing agent for exactly this kind of institutional need: an AI support assistant that connects directly to an institution's existing knowledge sources and platforms, answering questions across the student and staff lifecycle, with built-in guardrails to eliminate hallucination risk. It's available as part of the LearnWise AI platform alongside AI Student Tutor, AI Feedback & Grader, and AI Ops Assistant.

Deploying AI Campus Support with LearnWise AI

IE's initial goal was straightforward: one assistant, serving information for all of its programs, available to students around the clock. The team started narrow, testing AI Campus Support as a student-facing tutor and 24/7 assistant hosted on its website, connected directly to existing content. It quickly became clear that a single assistant couldn't sensibly cover every program: with many programs and many intakes, each needing its own rules and its own highly personalized support from program management teams, one shared build meant too much information to divide sensibly in one place.

Once the team recognized a single assistant wouldn't work, it tried deploying multiple agents instead, each covering a different program. That's when deployment hit a wall in Blackboard: there was no way to add a floating button or run multiple distinct agents the way IE wanted. The team pivoted to an assistant-only build hosted on its website, dropping the tutor functionality and keeping each agent focused on answering questions for its own program. That approach ran more reliably, but scaling it across the institution still proved hard. Centralizing content under one small team meant that team would have had to manage and keep accurate every program's content on behalf of every other team, which wasn't sustainable as the workload grew in size and complexity.

The problem wasn't the tool, the team eventually realized. It was the attempt to centralize. Pulling every team's fragmented knowledge, each managed in its own place and its own way, into one shared knowledge base meant asking people to change how they already worked, which created more friction than it solved. As Janowska put it: "We realized that we were just trying to battle it from the wrong angle. We were trying to put it all in one."

Instead of centralizing the knowledge base, IE left each team's existing sources exactly where they were, and let LearnWise become the single point of access across all of them, without disrupting any team's day-to-day work. Giving teams a tool to navigate their own sources, rather than asking them to clean up and hand that content over to a central system, solved their actual problem instead of creating new work for them. That's what let real ownership take hold, and what started the organic growth of new assistants across the institution. The result, in Janowska's words: "People started having ownership of their own content. It was so easy to learn how to use it, so that was great." Today, no dedicated team is needed to run the system: one person manages accounts, with LearnWise's technical support handling the rest.

What surprised the team most was how quickly the tool spread to internal, staff-facing use. Suppliers and Finance, and separately Talent & Careers, built their own assistants for onboarding new hires and short-term staff, without being asked to. As Janowska described it: "We were focusing so much on LearnWise for learning or for university cases or for students or for faculty, and all of a sudden it was such a clear corporate case that just grew naturally."

Decentralization also allowed every team to build flows specific to their own program rather than forcing every intake into the same logic. Those flows draw on Microsoft Entra data to identify students by name and surname, which helps the system recognize who's asking and route them to more relevant, personalized content. At IE's request, that identification has since been updated to run on anonymized data instead, which Janowska says has simplified the legal side without losing the system's ability to recognize a student and personalize their experience. 

On what stood out about working with LearnWise specifically, Janowska was direct: "I'm loving working with [LearnWise AI], and how quickly the responses are, how open [they] are, and how [they] understand the needs of the institution. It's not like [they] put in a ticket and it's just over there. This collaboration feels like real collaboration, like working together and building it together."

Results

Since adopting the federated model, IE has launched four assistants: a support assistant for its MIMS and MIM master's programs, KnowIE for Staff, KnowIE for Master Students, and a Talent & Careers assistant. Together, this portfolio has handled 1,730 conversations and 3,886 student and staff messages since December 2024, with a 94% resolution rate led by the Talent & Careers assistant. Thematic analysis of student questions shows four main categories: 

1. Career support (jobs, internships, CVs, interviews, the Career Portal),
2. Academic logistics (electives, exams, graduation, attendance, labs), 
3. Tech and platform support (Canvas, Blackboard, WiFi, portal access), and 
4. Campus services (library, parking, student card, wellbeing routing). 

IE also flags where gaps remain: operational step-by-step guidance such as academic calendars, IT ticket visibility, internal career processes, and faculty grading exceptions is still a weak spot, and wellbeing-related questions have surfaced often enough that sensitive-topic routing needs to be explicit rather than assumed.

The number of assistants has grown past those original four, with new ones launching as more teams opt in.

How does IE University govern AI Campus Support across different teams?

Growth has been organic and peer-driven ever since, with teams hearing about the tool from colleagues and asking to be onboarded rather than being pushed to adopt it from the top down. As Janowska puts it: "When you show the users at your institution that it's actually working, and it's easy, they will start knocking on your door. You won't have to push it." Rather than fearing it would replace them, Janowska says the reaction has run the other way: "It's more of 'Oh gosh, I hope it will replace me for all the boring stuff.'" One concrete example is the EdTech Assistant, built into IE's internal EdTech Scouting Platform, which helps faculty search, evaluate, and share teaching tools without digging through scattered resources.

Governance has stayed deliberately light. IE's central team retains visibility across all agents, but day-to-day content ownership sits with individual teams: two business school programs, a team supporting professors, Talent & Careers, Suppliers and Finance, and separately, a large Canvas-based team running its own B2B rollout. The central team periodically reviews usage to decide whether an agent is active, needs updating, or should be retired. Growth has been organic and peer-driven ever since, with teams hearing about the tool from colleagues and asking to be onboarded rather than being pushed to adopt it from the top down.

Looking Ahead

Starting in September, agent management is handing over to IE's IT staff, alongside an internal communication campaign to drive further adoption. The team is also exploring new integrations beyond its current toolset: an earlier attempt to connect with ServiceNow didn't fully work due to role complexity on IE's end. Other ideas in progress include a mentoring tool that gives temporary access to course materials for a limited window after a course ends, and a tailor-made project the Lifelong Learning team is exploring separately.

See how AI Campus Support can bring the same kind of ownership and scale to your institution's student and staff support. Book a demo.

This piece is adapted from IE University's session at ICDLE 2026, "From Pilot to Practice: Building an Institution-Wide AI Support Strategy at IE University," and a follow-up conversation with LearnWise AI. As ICDLE does not make breakout session recordings publicly available, this case study is built from the presenter's own materials and a direct interview rather than session video.