Should a Community College Automate Advising Triage or Cut the Number of Advising Paths?

The question you were asked, and the one you were sold

Your vice president of student success came back from a conference with a demo. The vendor opened on the same slide every vendor opens on: advising caseloads at community colleges run as high as 1,200 students per advisor, a figure the Community College Research Center has been publishing since 2021. The pitch that follows is predictable. A model scores every student for risk, routes the top of the list to an advisor, and the advisor finally has a manageable day.

The demo is honest about what it does. It makes 1,200 to 1 survivable. What it does not do, and what nobody in the sales cycle will price for you, is ask whether 1,200 to 1 is a staffing shortage or a design choice.

At most colleges we have looked at, it is mostly a design choice. Not a deliberate one. It accreted.

Where the caseload actually comes from

An advisor's day is not consumed by 1,200 students. It is consumed by the number of distinct answers those students can need. A student in a nursing pathway with one sequence, one set of prerequisites, and one transfer destination is a five-minute conversation. A student choosing among four associate degrees in adjacent business fields, each with a different math requirement, two of which articulate to the regional university and two of which do not, is a forty-minute conversation that has to be repeated next term because the student changed their mind.

Multiply that second student across a catalogue. A college with sixty degree and certificate options is not unusual. A college where a large share of those options enroll fewer than twenty students each is also not unusual, and every one of them carries its own map, its own course sequence, its own exceptions, and its own set of questions an advisor has to be able to answer on demand.

That is the caseload. The headcount is the numerator. The catalogue is the multiplier. Vendors sell to the numerator because that is where the software goes.

Two interventions that are substitutes, sold as if they were not

There are two things a college can do about advising load, and they compete with each other for the same budget and the same year of leadership attention.

The first is triage automation: early-alert, predictive analytics, caseload management, whatever the vendor calls it this quarter. It takes the complexity as given and helps advisors prioritise within it. It is a coping mechanism, and a good one.

The second is path reduction. Guided Pathways, the reform the sector has been pursuing for a decade, is explicitly a structural intervention: fewer, clearer programs organised into meta-majors, default course sequences, and advising organised by field rather than by alphabet. It is not a tooling reform. It reduces the number of distinct answers an advisor has to hold, which reduces the load without adding a single seat.

These two are substitutes at the margin. Every program you retire removes a map from the advising queue permanently. Every triage license you buy removes nothing and instead makes the existing maps tolerable, which is a polite way of saying it locks them in place. A college that buys early-alert software before rationalising its catalogue has just made the strongest possible argument against ever rationalising it, because the pain that would have forced the conversation is now managed.

We have never seen a vendor deck compare the two. We have rarely seen a college compare them either, because they live in different budgets. Software is a purchase. Program review is a governance fight.

What the evidence actually says about path reduction

The honest version of the research is more useful than the promotional version, so here it is.

CCRC's March 2024 working paper by Veronica Minaya and Nicolas Acevedo examined 62 community and technical colleges across Tennessee, Ohio and Washington and fourteen distinct guided pathways practices. Tennessee showed significant positive associations with early outcomes: fall-to-fall persistence and first-year credit accumulation. Ohio and Washington did not. The authors are careful to say they cannot establish causation, and they are equally careful about the finding that matters most for a buyer: the gains came from complementarities among practices, not from any single practice or from how intensely one was adopted.

Read that as a warning against buying the software version of one practice. An early-alert tool is, functionally, a single practice adopted intensely.

The more recent evidence points the same way. CCRC and the Aspen Institute reported in 2026 on ten colleges in their Unlocking Opportunity network, which launched in 2023. Between 2022 and 2025 those colleges moved 20,104 more students into what the report calls high-value pathways and cut enrollment in low-opportunity programs by roughly a quarter, 7,648 students. Odessa College in Texas eliminated its general studies associate degrees outright because the post-graduation prospects did not justify them. That is path reduction in its plainest form, and the colleges that did it also redesigned onboarding and advising around the smaller catalogue.

Notice what is missing from those results. There is no line for software spend. The reforms were mostly funded, as Davis Jenkins of CCRC put it in an October 2025 interview, by redeploying existing resources, and he added that there is a limit to how much community colleges can keep doing more with less. That limit is real. It is also the strongest argument for spending the next dollar on removing work rather than on managing it.

The pressure to buy is going up, not down

Tyton Partners' Driving Toward a Degree 2025 report, drawn from surveys of more than 3,000 administrators, advisors and students, found that staffing shortages are pushing more than 70 percent of large and public institutions to increase advisor caseloads, even though high caseloads are the barrier those same institutions name first. Nearly 40 percent of four-year publics expect cuts to student support budgets over the next three years.

That is the environment in which the demo arrives. Caseloads rising, budgets falling, a board asking what you are doing about retention. Triage software is the answer that can be procured in a quarter. Program rationalisation takes a year of curriculum committee meetings and makes enemies in three departments. We understand why the purchase order wins.

But the purchase order is answering a question about this year's pain, and the catalogue is a question about the next decade's cost structure. The college that automates triage in 2026 will still have sixty programs in 2031, plus a renewal invoice.

The test we run before either answer

We do not tell a college to skip triage software. Some colleges genuinely have a lean catalogue and a caseload problem, and for them the tool is the right buy. The question is which kind of college you are, and it has an answerable form.

Take your advising contact logs for one term and code each contact by the program decision it involved. Not the student's declared major. The decision: which of several adjacent programs, which math sequence, which transfer destination, whether a course counts. Then look at the distribution of those decisions across the catalogue.

If the load is spread across most of your programs, you have a caseload problem and triage will help. If the load concentrates in a small number of program clusters where students bounce between near-identical options, you have a path problem, and the software will make it comfortable rather than fix it.

Most colleges we have seen fall into the second group and buy for the first. The catalogue has a lot of degrees of freedom, and none of them are free.

What this costs to get wrong

The visible cost of choosing triage over path reduction is the license, which is modest. The invisible cost is that you have removed the pressure that would have forced the harder decision, and you have done it at the moment, with 70 percent of institutions raising caseloads, when that pressure was finally high enough to act on.

The reverse mistake exists too. A college that cuts programs without redesigning onboarding and advising around the smaller catalogue gets the political cost of the cuts and none of the operational benefit. The Tennessee result was not fewer programs alone. It was fewer programs plus the practices that depend on there being fewer programs.

That is the same pattern we describe in what districts spend on enrollment paperwork: the process cost lives in work that no budget line names, and software gets bought against a number nobody has actually measured. It is also a specific case of the general rule in when not to use AI in business, which is that automating a process you should have simplified freezes the complexity in place. And it is why we tell higher-education clients that the ROI of an AI investment depends less on the model than on what was true of the operation before the model arrived.

The honest edge of what an article can tell you

We can tell you the two options are substitutes and that almost nobody prices them against each other. We can tell you how to code a term of advising contacts to find out which problem you have. We cannot tell you, from here, what your distribution looks like, which forty of your sixty programs are carrying the advising load, or whether your curriculum governance can survive the conversation that follows. That takes a week inside your contact logs, your catalogue, and your committee minutes.

If your college is holding a triage software quote and has not yet looked at whether the caseload is a staffing number or a catalogue number, that is the conversation to have first.