Your cabinet just signed off on a strategy paper. On page four, it says the institution will be a "University 5.0" campus by 2030. Human-centric. AI-enabled. Working closely with industry and the community.
Last month, your team finished a discovery audit. It found 61 tools running across academic and administrative departments. You had approved nine of them.
Same institution. Two very different pictures.
That gap is worth sitting with, because it is not really an IT problem. It is a problem that looks different depending on which chair you were sitting in when the paper went round.
University 5.0 is a model of the university built around people rather than output, where AI and data are used to serve students, staff, and the wider community instead of only making the institution more efficient.
The idea comes from Industry 5.0. The European Commission described it in January 2021 as an industry that is sustainable, human-centric, and resilient. Japan's Society 5.0 is the other parent. On a campus, it usually shows up as four commitments. Personalized learning. Student and staff wellbeing. Sustainability. Closer work with the industry, government, and the local community.
University 4.0 was the connected campus. Cloud SIS, learning analytics, mobile services, data platforms. Technology is brought to deliver services at scale. University 5.0 keeps all of that and adds one condition: the technology has to serve a human outcome, and someone has to be able to show that it does.
That last clause is where it gets hard. Showing it needs a record. Most institutions cannot produce one.
Think about how each version gets written up.
University 3.0 was the entrepreneurial campus. Tech transfer, spinouts, research income. University 4.0 was the connected campus. University 5.0 is human-centric AI, wellbeing, and sustainability.
Every one of those is described in two ways. What the student sees. What the institution buys. None of them says anything about how the work gets done between offices.
That matters more than it sounds. Your Banner or Workday go-live probably went fine. The system does what it was built to do. But the change the business case promised never quite showed up, because systems of record hold records. They do not handle the work that crosses between them. The appeal. The exception. The committee decision. The request that touches four offices and belongs to none of them. That work still runs on email, shared drives, and whatever tool a department paid for on a card.
EDUCAUSE said this out loud in the 2026 Top 10, published in October 2025 under the theme "Making Connections." The report describes higher education as largely siloed across data, technology deployments, and even strategic planning and decision-making. It adds that institutions are sitting on valuable data while each digs its own separate tunnel toward it.
That is the gap 4.0 skipped. It is where 5.0 gets decided.
Shadow IT used to move slowly. A department bought a tool. You found it either 11 months later at renewal or during a security review. Slow, but slow enough to catch.
AI does not work like that. Someone can pick up a new tool before lunch.
EDUCAUSE published a study on AI and work in higher education on 12 January 2026. In the previous six months, 94 percent of the people surveyed had used AI for work. Almost all of them, 92 percent, said their institution has an AI strategy. But only 54 percent knew what their institution's AI policy actually said. And 56 percent were using AI tools that their institution had not approved.
Sit with those last two numbers.
More than half of your colleagues are doing university work on tools nobody signed off. Roughly the same number do not know the rules. And the two things they most wanted AI to help with were repetitive tasks and admin burden, so the pull is coming from the back office rather than the classroom.
The strategy paper and the discovery audit do not describe two institutions. The audit shows what the strategy looks like with nothing underneath it.
It gets worse. That same study found only 13 percent of institutions measure what they get back from AI. If you cannot measure it, you keep funding the visible things. Pilots and platforms get money. The plumbing does not.
Here is where one document becomes eight problems. I have put them roughly in order of how badly each underlying job is served today, rather than by seniority, which puts the Provost first and the CIO third.
The Provost is looking at an accreditation cycle. Every AI-supported teaching initiative in that paper is an academic decision, and academic decisions need a record. Right now, the record of who approved what, under which policy version, sits in email threads across six offices. When an accreditor asks how a curriculum decision was reached eighteen months ago, the answer should not involve searching mailboxes. Of everything on this list, this is the job currently served the worst.
Research administration and the IRB read the same paper and see the queue length. AI in research triggers more ethics review, not less. Human subjects, data provenance, and model use in analysis. Ethics review is already the longest leg of a grant submission timeline. University 5.0 makes it longer, and a missed submission deadline carries a number.
The CIO has to make "AI woven through the institution" safe. More than half of the staff are already working in tools that nobody approved, and the discovery audit is now a month old and already out of date. You cannot audit your way back to control when the thing you are auditing arrives faster than the audit cycle. This is the chair most people assume the paper is addressed to, and it is not even the worst-off one.
The Registrar, Admissions, and Financial Aid hear a promise about a personalized student experience. Personalization is made at the front door and broken in the handoffs. A transfer-credit evaluation that takes three weeks is not personalized by any model you put in front of it. The applicant has already been accepted somewhere faster.
The CISO is being asked to sign off on a category rather than a system. Staff is already moving institutional data through tools that never came under review. The policy exists. Only about half the people know what it says. That is a hard position to sign from.
The CFO sees a number with no line item. Only 13 percent of institutions measure what they get back from AI, which means the sector is asking finance to fund a category whose return nobody is yet measuring, during a revenue squeeze. That is not an obstruction. That is the job.
Procurement and operations are about to receive dozens of pilots at once. Every one is a purchase that falls outside standard ERP rules, and every one needs a security review, a data agreement, and a renewal date somebody owns. Multiply that by every department that reads the paper.
HR gets the word "human-centric" and has to turn it into a workforce plan. The institution just committed to reskilling, and 69 percent of institutions building AI capability are upskilling existing staff rather than hiring. That is a credentialing and records workflow running across systems that do not talk to each other.
Eight readers. Eight different Mondays. One sentence on page four.
Whatever lands on your desk with a 5.0 label attached, these three will tell you quickly whether it has a delivery plan or just a vision.
When this needs a decision from two or more offices, where does that decision happen, and who writes it down?
If an auditor or an accreditor asks how a decision was made eighteen months ago, does someone have to go digging through a mailbox?
If a department wants to change its own process, can it do that without raising a ticket, and can your team still see what changed?
A project that fails all three ends up where 4.0 ended up. A system that works. A vendor who is happy. And a pile of workarounds nobody planned, and nobody governs.
What all eight chairs are describing is the same missing thing. Call it a governed execution layer: the place where work that crosses departments actually runs, under the access rules, audit standards, and identity setup IT already manages.
It sits next to Banner, Workday, Ellucian, or PeopleSoft. Those systems keep owning what they own. The layer holds what falls between them. It routes a decision to the right owner instead of forwarding an email. It records who decided, when, and under which version of the policy, as a by-product of the work rather than a document assembled afterward. It holds the exception, which is where the institution actually spends its time. And it lets the person who owns a process change it without opening a ticket.
Kissflow is a layer for work at the edges of the SIS and ERP. Departments build the apps and workflows they need, and those apps run inside the governance the IT team already runs: single sign-on, role-based access, and an audit log that stamps and names every action.
The part that matters most here is what the AI hands back. Kissflow's AI Builder has been generally available since May 2026. You describe a process in plain language, and it returns a blueprint first: the data model, the pages, the roles, the workflow, and the navigation. It does not hand you application code to maintain. So the logic stays readable for a compliance officer or an accreditor, the app stays editable by the person who owns the process, and changes made by AI write to the same audit log as changes made by hand. A person stays in the lead.
Kissflow holds SOC 2 Type II, ISO/IEC 27001, HIPAA, and GDPR compliance, supports single sign-on through SAML and OAuth, and connects to the systems you already run. Over 1,200 customers use it worldwide.
Neither needs a purchase order, and both are worth doing, whatever you end up buying.
Take that discovery audit and sort it differently. Not by department, not by spend. Sort it by the process each tool is covering, then mark every process that crosses more than one office. That second list is your real gap, and it is usually shorter than the tool count makes it look.
Then pick the three busiest crossings on it and ask each office one question: where does this get stuck, and who has to chase it. You will have an answer inside a week, and it will name a handoff rather than a system.
Take both lists to the cabinet alongside the 5.0 paper, and present them as the delivery plan the strategy is missing.
The institutions that make 5.0 real will not be the ones that bought the most AI. They will be the ones who can move a decision across four offices, write it down, and produce the evidence eighteen months later without anyone opening a mailbox. That is a governance job before it is a technology job, and it is the one your board will eventually ask you about.
Book a walkthrough of how Kissflow governs cross-department approvals alongside your SIS and ERP without adding another system of record.