ChatGPT
Claude
Perplexity
Gemini
GrokSummer melt is a US term, and Canada has the same summer gap
"Summer melt" comes from US research on students who intend to go to college, often with a deposit paid, and never arrive in the fall. Canadian institutions face the same gap: a student accepts an offer in the spring, then does not show up in September, or switches to another institution over the summer.
The published trials on how to reduce it are American. We say that up front because none of the numbers below measure Canadian students, and we have not found a published Canadian trial of an AI assistant in this window. What follows is what the US evidence supports, and how to adapt it to Canadian admissions, where the process differs by province and by institution.
A note on method: no paid keyword-volume tool was available for this article, so it was scoped from public search results. Every figure below comes from a source we opened. Nothing comes from Skolbot or its customers.
What the US trials found
Castleman and Page estimated US summer melt "ranging from 8 to 40 percent" depending on the population. The wide range is the point: it varies by who the students are, and you cannot import it as a Canadian rate.
| Study | Design | Published result |
|---|---|---|
| Castleman and Page, "Summer Nudging", 2015 | Automated, personalized texts or near-age peer mentors over the summer | Gains among students with less access to counselling during the school year |
| What Works Clearinghouse review of the same study | 6,196 students, ten texts over July and August | Most comparisons not statistically significant; clearest gain in two Massachusetts sites; none in Boston |
| Page and Gehlbach, AERA Open, 2017 | Conversational AI texting incoming first-year students at Georgia State University | Committed students were "3.3 percentage points more likely to enroll on time" |
Both interventions were short, tied to a task the student could complete, and timed to a deadline. Neither tried to change a student's mind about the institution. The Georgia State assistant also predates generative AI, so treat it as proof of a pattern, not a forecast. Our reading of the wider research on admissions chatbots covers the other studies and their limits.
How the US setup differs from Canada
| United States | Canada | |
|---|---|---|
| Commitment | Deposit, typically by May 1 | Offer acceptance in the spring; in Ontario, universities generally cannot require a reply before June 1 |
| Application route | Common App or direct application | Varies by province: central services in Ontario (including ontariocolleges.ca for colleges), direct applications elsewhere |
| Funding | Federal aid forms, institutional awards | Provincial student aid plus federal student financial assistance, and institutional awards |
| Summer window | May to August | Roughly June to early September |
The consequence is that the Canadian summer window looks a lot like the US one: a few months between commitment and the first day of class. The tasks are different, though. Provincial aid applications, residence, transcripts and course registration each create a step where a student can stall.
Where an AI agent helps an admissions team
An agent can answer a student's question late on a Sunday: "Why can't I register for courses?", "Where do I send my final transcript?", "Has my residence deposit been received?" It can also send reminders tied to the student's own checklist. Those two jobs map onto the Georgia State design.
An agent does not decide aid, waive a hold, or talk a family through a financial crisis. Those go to a person. Our definition of an AI admissions agent sets out what an agent answers and what it hands off.
| Job | Agent | Staff member |
|---|---|---|
| Reminder before a residence or registration date | Yes | Not needed |
| "Where do I send my final transcript?" | Yes, from approved content | Backup |
| Questions about a provincial aid assessment | Gives the general process, then hands off | Yes |
| A student considering another institution or a gap year | Flags and routes quickly | Yes |
| Confirming that a deposit or step is complete | Only if connected to the student record | Backup |
The last row is a capability question. An agent that cannot see the student record can only give general answers. Ask any vendor, including us, which systems the agent reads and writes, and test it before signing.
Who is most at risk in a Canadian cycle
The US trials found their clearest gains among students with the least access to advice. That is a useful lens for a Canadian team. First-generation students, students moving from another province or a rural area, and students who depend on provincial aid often have no one to ask the small practical question that decides whether they arrive in September.
Your own data will tell you more than any study: look at who accepted an offer last year and never registered for a course.
A summer plan to test
This is an illustrative workflow, not a measured result. It follows the pattern the trials share.
- Build the checklist per student. Start from the steps your own data show hold up enrolment: aid, residence, final transcript, course registration, fees.
- Send fewer, dated messages. The Castleman and Page design used ten texts over two months. More is not better.
- Make every message answerable. A student who replies should get a useful answer, not a no-reply notice.
- Route emotion to people. Messages about cost, family pressure or doubt should reach an advisor the same day.
- Prioritize the students with the least backing. The trial gains concentrated among students with fewer other sources of help.
- Clean the contact data first. A message sent to a wrong or missing number reaches no one, and the dashboard will not tell you.
Measure with a control group
Hold back a random sample of admitted students from the new outreach and compare the two groups on one outcome, such as course registration by a set date or attendance in the first week. Year-over-year comparisons mislead because aid rules, housing supply and the economy change from year to year.
Privacy and consent
Which privacy law applies depends on your province and on the type of institution. Public universities and colleges are generally covered by provincial public-sector privacy laws, and the Office of the Privacy Commissioner of Canada explains how the federal and provincial regimes fit together. Confirm with your privacy office which rules cover admitted students before you connect an agent to your records.
Electronic messages can also fall under Canada's anti-spam legislation. Whether an admissions reminder counts as a commercial message is a question for counsel. Tell students plainly that they are talking to an automated assistant, and keep admission decisions with people.
Where this fits in your recruitment plan
Summer melt is the last stretch of a longer funnel. The guide to AI agents for student recruitment shows where a post-offer agent sits in the full journey, from first inquiry to the first day of class.
