Some of your institution's most important work happens once a year. Often only one person remembers how.
Colleges and universities run on lean offices, long tenures, and annual cycles. That combination keeps critical knowledge in a few people's heads, where it is hard to hand off, hard to find, and unavailable to the AI tools your staff and students now use.
Why Institutional Knowledge Is Fragile in Higher Education
Institutional knowledge in higher education is fragile because the people who hold it rarely have a backup, and the work that depends on it often comes around only once a term, once a year, or once a decade. Administrative offices tend to run lean, so a single person in financial aid, the registrar's office, or institutional research may be the only one who knows how a rule is applied or why a report is built the way it is. Long tenures make that knowledge feel like second nature to the person who holds it, which is exactly why it rarely gets written down. And because divisions and academic departments operate with considerable independence, what one office knows often never reaches the others.
Core Issues
- Single points of failure. One person holds the working rules for a process that the whole institution depends on, and no one else has written them down.
- Knowledge lost between cycles. Work that happens once a year, or once an accreditation cycle, depends on someone remembering how it was done last time.
- Leadership transitions. New presidents, provosts, deans, and vice presidents inherit divisions that each hold part of the picture, along with years of decisions they weren't present for.
- Numbers no one can fully explain. Enrollment, retention, and aid figures can differ depending on who ran the report, because the definitions behind them live with the people who built them.
- Uneven onboarding. Beyond HR orientation, each office teaches new staff its own systems and procedures, and quality depends on who has time to teach.
- Scattered systems. Institutional data sits across the student information system, the learning platform, the CRM, the advancement database, and departmental spreadsheets that leadership may not know exist.
The Calendar Decides When Knowledge Gaps Show Up
In most organizations, a missing piece of knowledge surfaces within days. In a college or university, it can stay hidden for months, until the aid packaging cycle opens, the budget build begins, or the accreditation self-study that comes around once every eight to ten years is due. By then, the person who handled it last may have retired, moved to another institution, or taken on a different role.
The once-a-cycle task
A process that runs annually is re-learned every year from old email, shared-drive folders, and whoever happens to remember, and small errors in how rules are applied can carry through for students.
The inherited decision
A new leader asks why a policy, program, or reporting method was set up the way it was. The reasoning exists somewhere in years of committee and cabinet minutes, but no one can find it.
The system migration
Moving to a new student information system or ERP forces the institution to decide which custom codes, calculations, and workarounds to carry forward, often before anyone has documented what they do.
AI Answers With Whatever Your Institution Has Written Down
Colleges and universities are putting AI to work in admissions, advising, IT help desks, and everyday administrative tasks. An AI tool answering a student's question about an aid deadline or a transfer credit is only as reliable as the policies, catalog pages, and procedures it draws on. Where those sources conflict, are out of date, or were never written down, AI doesn't resolve the problem. It repeats it, confidently and at scale.
The institutions that get useful results from AI and self-service tend to have one thing in common: people who know the answers did the work of writing them down first.
Georgia State's enrollment assistant worked because admissions counselors wrote its first answers.
Counselors seeded an AI text-message assistant with about 250 frequently asked questions. When it couldn't answer, questions went to staff, and their replies were added back until the knowledge base grew past 1,000 answers. In a randomized trial, committed students who received it were 3.3 percentage points more likely to enroll on time, a 21% reduction in summer melt.
Page & Gehlbach, AERA Open, 2017Griffith University tripled IT self-service after expanding its knowledge base.
Students and staff relied on service personnel for routine requests. As part of a broader service-management rollout, Griffith built out more than 1,300 knowledge articles and integrated digital services. Self-service rose from 21% to 63%, while calls fell 31%, emails 46%, and walk-ins 26%.
ServiceNow, Griffith University customer storyA retiring University of Michigan leader's tacit expertise was captured in four structured interviews.
A senior development leader was retiring with judgment and experience no handover checklist could hold. Researchers used four structured interviews to retrieve, validate, and document that knowledge for incoming leadership. Two years later, the organization was still using it in recruiting, mentoring, coaching, and training.
Journal of Knowledge Management, "Leadership transitions, tacit knowledge sharing and organizational generativity"How Glymr Helps
Glymr works with administrative and academic leadership to make institutional knowledge usable for the staff, leaders, and AI tools that depend on it. Work can start with a single office, a division, a leadership transition, or a system migration.
Protect continuity. Critical Knowledge Capture documents what long-tenured staff know about the processes, exceptions, and history that no procedure manual covers, before a retirement or departure turns it into a gap. Knowledge Exit Interviews give departing staff a structured way to hand that knowledge over. For a president, provost, or vice president stepping down, Executive Knowledge Transfer is designed to give a successor the context, relationships, and reasoning behind past decisions that a transition memo leaves out.
Make systems and processes visible. Data Landscape Mapping identifies the systems, databases, and spreadsheets that hold institutional data, including ones leadership doesn't know are still in use, which is especially valuable before a student information system or ERP migration. Process Mapping documents how cross-office work such as financial aid, registration, and transfer credit evaluation actually runs, including the exceptions that determine how students are affected.
Keep knowledge current and findable. Knowledge Management Programs put ownership, review cycles, and a clear structure behind policies, procedures, onboarding materials, and data definitions, so they stay accurate from one cycle to the next and give AI tools reliable context to work from.
Where to Start
The two assessments answer different questions.
A Knowledge Friction Assessment is the right starting point when the concern is continuity and everyday work: staff searching for information or waiting on colleagues, repeated questions, uneven onboarding, and critical knowledge concentrated in too few people. It measures the friction that affects how your offices run today.
An AI Effectiveness Assessment fits institutions where faculty and staff are already using AI regularly. Based on your own survey results, it shows whether that use is turning into useful work and where it breaks down, whether in the knowledge AI draws on, in access and governance, or in the time staff spend checking and correcting output.
Who This Is For
This is most relevant for presidents, provosts, and administrative leaders at colleges and universities facing retirements or staff turnover, a leadership transition, a student information system or ERP migration, or a rollout of AI tools across offices that rely on institutional knowledge few people hold.
Related Reading
See Critical Knowledge Risk for a closer look at what happens when essential knowledge is concentrated in a small number of people.
Find out which parts of your institution depend on knowledge only a few people hold.
A Knowledge Friction Assessment shows where staff lose time and where critical knowledge is concentrated. If AI is already in wide use across your offices, an AI Effectiveness Assessment shows whether that use is turning into useful work.