Reactivating inactive students: AI campaigns for training schools
Every school has a base of past students and cold leads that never converted. It’s the richest and most forgotten CRM in the sector. How to reactivate without sounding desperate.
Every training school over three years old has — in the CRM or in spreadsheets — a huge base of "dormant" contacts: past students who did a course and never came back, leads who asked for info and didn’t convert, participants at free workshops who never signed up for a paid one.
It’s the richest and least-used base in the sector. And it’s where a well-designed campaign generates far more value than spending on ads.
Why traditional reactivation fails
Three predictable reasons:
- Generic. "New courses available!" to everyone. No signal the school remembers who the person is.
- Poorly timed. Arrives when the person isn’t thinking about training, and doesn’t arrive when they are.
- Single channel. Only email, or only SMS. Loses half the audience each attempt.
What AI changes in a reactivation campaign
- Segments by course taken, previous interest level, and recent signals (site visit, social interactions).
- Writes in the first person — as the coordinator of the course the person took —, which triples reply rates versus brand emails.
- Picks the channel per profile: WhatsApp for those who interacted on WhatsApp before; email for email-preferrers; SMS for no-app numbers.
- Signal-based timing — sends when the person visited the site in the last week, not on the first of the month.
Three approaches that work (and one that doesn’t)
Works:
- "Almost done": for those who started a course and didn’t finish, with an adapted completion plan.
- "The next step": for those who completed a course, propose the advanced one — with a real loyalty discount (10-15%).
- "Your field changed": for former students in fast-moving areas (data, AI, digital marketing), offer a free 60-min session on what’s changed.
Doesn’t work:
- "Special prices today just for you!" — sounds like door-to-door sales. Whoever said no once won’t change their mind because of 10%.
Realistic numbers
On a base of 3-5k dormant contacts, a well-designed campaign generates:
- 8-12% active response (click, reply, site visit).
- 1-3% enrolment conversion — significantly more than cold ads.
- CAC 4-6× lower than Google Ads campaigns in the same period.
Where AI shouldn’t go
- Students with logged complaints — don’t touch them, and flag them clearly in the base to exclude.
- Former students who asked not to be contacted — GDPR, always.
- People who lost someone close (if you know). A music or arts school has these situations; a blind automatism is indefensible.
The well-run campaign is two weeks of setup and one month of calibration. From then, it’s a machine the school runs once a quarter — and that returns money already sitting there.