A Third of Entry-Level Jobs Now Require AI Skills. A New Degree Program Takes 18 Months.

This article was originally published in Entrepreneur on September 7, 2026. It is the first piece in a series called The University Intrapreneur.

Every student walking across a stage this spring is being measured against a hiring bar that moved twice since they declared their major.

More than a third of entry-level jobs now ask for AI skills, according to the National Association of Colleges and Employers, nearly triple the share reported in fall 2025. That change took two semesters. One hiring cycle. The length of a single academic year.

I have spent 20 years inside innovation programs at companies and universities on six continents, and I have watched the same thing happen every time an institution waited for permission that was never going to arrive.

What the institution's clock actually looks like

A new degree program can take up to 18 months to move from proposal to approval. A tenure-track faculty search runs 9 to 12 months from posting to a start date. Curriculum at many institutions comes up for review on a seven-year cycle. Those timelines protect quality, and they were built for a world where the job a graduate walked into looked the same at the end of the process as it did at the start.

Policy moves at the speed the institution already moves. People move faster. The ones who close this distance are already inside the building, working while the slow vote is still being counted.

Who drives change inside a university

The university intrapreneur can be anybody. The role usually lands on a mid-level faculty or staff member who found a real problem and built a working answer on borrowed time. It carries no fixed title and no fixed department. The org chart does not account for that person, the budget has no line for the work, and the institution's ability to change depends on them anyway.

Gifford Pinchot III coined the word intrapreneur in a 1978 white paper written with Elizabeth Pinchot, and called them dreamers who do. Pinchot was writing about corporations. The university version carries more weight now, because the distance they absorb is wider.

Faculty, staff, a department administrator, someone in the library. Most of them already have the talent. What they need is a name for the role, a tool built for the problem in front of them, and an institution willing to give support and then step out of the way.

Universities have moved this fast before

The GI Bill put about 8 million veterans through education, and by 1947 half of American college students were veterans. A public program and a few thousand campuses absorbed a generation in under a decade.

UTeach is the smaller version of the same instinct. It started as one faculty-built route at the University of Texas at Austin in 1997, and it now runs at more than 40 universities. Nobody wrote it into a plan first.

What one person can do before the committee meets

1. Name the role, then name the people already doing it, where others can hear it. Nobody volunteers for something with no name, and nobody keeps building something nobody has celebrated. Look for the unofficial tool half a department relies on, the pilot that outlived its funding, the student club that solved something the provost's office is still studying. Who on your campus is already doing this work without a title for it?

2. Build the channel before you build the lab. One shared, visible place, open to faculty, staff and students, where builders say what they are working on before it is finished. The campus that talks to itself catches two departments building the same project in parallel before either one runs out of runway. Keep the bar to enter low, and give whoever uses it air cover: a named senior sponsor, and one sentence you will repeat unchanged when someone objects. Where would that conversation happen at your institution today?

3. Put money where the experiment is. Permission is free, and money is what turns permission into a pilot. AI has cut the cost of standing up a pilot low enough that a department can fund one from its own discretionary budget. Name a person who can approve that spend without a committee, and put a decision date on the calendar beside their name. A review date lets everyone postpone without feeling like they said no. A name attached to a date forces a yes or a no. Who holds that authority where you work, and do they know they hold it?

What adoption looks like when nobody assigned it

I advise the AI working group at the California State University system, the largest public university system in the country, with 22 universities and more than 471,000 students. CSU surveyed its own community in fall 2025 and heard back from more than 94,000 students, faculty and staff, the largest study of its kind in AI in higher education.

95% of respondents had already used an AI tool. 82% of students called AI essential to their profession, and the same share worried about their job security.

Nobody assigned that adoption. It happened because individual people decided that waiting carried more risk than building ahead of policy. The same pattern shows up in the hiring data, where entry-level roles are being rewritten faster than any catalog can follow, and across the 3,900 American colleges running the same experiment at once.

I hear a fair objection from the sharpest AI skeptics on every campus, and I take it seriously. They tell me a strategy is more than a tool, and judgment is more than an adoption number. They are right about that. The question underneath is whether the institution around the tool can still change shape when the world outside it does.

The capability a degree certifies now has a half-life measured in years rather than decades, which is the subject of the book I am writing now, The Great Relearning. A university absorbs that pressure through a person, or it does not absorb it at all.

The people best positioned to rebuild a university's capability are already inside it, on the payroll, doing the work on borrowed time. Find them, connect them to each other, and give them cover. The institutions that come through this decade will be the ones that kept producing people willing to build before anyone told them to.

I am collecting these stories for the rest of this series: the builders nobody has named yet. If that is you, or you know who it is on your campus, I am easy to find.

Here is what makes Alex a credible voice on this topic: Alex advises the AI working group at the California State University system, the largest public university system in the country. Before that he spent 20 years establishing Cisco's global innovation centers across 14 countries, most of them on university campuses.

Looking for the builders already inside your institution? book a conversation →