What an AI agent adds to enquiry follow-up
An AI agent does not reinvent follow-up. It chooses which pre-approved message suits a prospective student at a given moment, based on what they have done, and it knows when to stop. The discipline around it is the same as ever: clear permission, a sensible cadence and an adviser ready to step in.
In Australia, admissions run through state-based centres (UAC, VTAC, QTAC, SATAC and TISC) as well as direct applications, and many students apply on an ATAR or an alternative pathway. A student who sent an enquiry in September, read a course page in October and went quiet while waiting for Year 12 results needs something different from one who just booked an open day. This guide covers the triggers worth automating, sequence design, the consent framework under the Spam Act and privacy law, and the limits to settle before launch.
A note on method: no paid keyword-volume tool was available, so this piece draws on public search results. It contains no performance statistics, and nothing in it comes from Skolbot or its customers. For email fundamentals, see our guide to email nurturing for student prospects.
Fixed drip versus event-driven agent
A fixed drip sends the same messages on day one, three and seven. An agent adds a decision layer: it reads context (course viewed, questions asked, application status) and selects from templates and rules your team has approved. Your staff own the rules; the agent assembles within them.
| Aspect | Fixed drip | Governed AI agent |
|---|---|---|
| Starts when | A contact joins a list | An event: page view, question, missing document |
| Content | One template | Approved template adapted to course and stage |
| Stops when | The drip ends | The student replies, unsubscribes or books a call |
| Main risk | Irrelevant message | Inaccurate or pushy message at scale |
| Human oversight | At design time | At design time and on sensitive cases |
This describes one way to organise the work, not a specific product. Both approaches depend on the person having agreed to hear from you.
Triggers worth automating
- Enquiry or course guide request. An immediate reply, then a useful follow-up a few days later: course structure, entry requirements, pathway options. See our brochure-request sequence.
- A question asked on your site. If the visitor asked about HECS-HELP eligibility or credit for prior study, the next message picks up that topic.
- Open day or info session registration. Confirmation, reminder, then a next-day note with a link to speak to an adviser.
- Incomplete application. A factual reminder of the missing item and your own closing date, with a named contact if something is blocking them.
- Extended silence. One re-engagement message, then stop. The full method is in our article on reactivating dormant student leads.
Leave alone: offer decisions, refusals and one-to-one conversations about fees or financial hardship.
Build rules that make the sequence stop
- A ceiling. Set a maximum number of messages per topic and per period across every channel. An email, an SMS and a chat message on the same day for the same reason is three contacts.
- Stop on reply. The moment the student answers, automation goes quiet and an adviser takes over.
- Stop on unsubscribe. A refusal ends marketing on the channels concerned, immediately and permanently.
Add a fourth for tertiary admissions: respect the student's calendar. Someone sitting Year 12 exams or waiting on offers has less attention than in a quiet week. Confirm current closing dates and offer rounds with the relevant admissions centre or your own institution rather than reusing last year's.
To decide who deserves a call rather than another message, use structured lead qualification.
Consent in Australia: the Spam Act's three rules
The Spam Act 2003 sets three rules for commercial electronic messages. Unsolicited commercial electronic messages must not be sent (section 16), they must include accurate sender information (section 17), and they must contain a functional unsubscribe facility (section 18) (Spam Act 2003, Part 2). Consent is the first rule, and the Act's Schedule 2 covers it in more detail, including how consent can be express or inferred. Do not assume an enquiry form creates consent for every kind of follow-up; state clearly at the point of enquiry what you will send.
Voice calls are covered separately by the Do Not Call Register rules, so check them before adding any outbound calling to a sequence. Privacy law applies on top: the Privacy Act 1988 and the Australian Privacy Principles govern how you collect and use the personal information that feeds your triggers, and the Office of the Australian Information Commissioner oversees them. If you use offshore agents or recruiters, check the obligations that come with them. Ask your privacy officer to review any profiling before launch. This article is not legal advice on your process.
Limits to set before you automate
| Risk | Example | Safeguard |
|---|---|---|
| Inaccurate information | An invented fee or entry requirement | Answers restricted to an approved knowledge base; hand-off when unsure |
| Implied promise | "You'll get an offer" or "last places" without basis | List of banned phrasings reviewed by admissions |
| Over-contact | Three channels, same day | One ceiling per person across channels |
| Opaque profiling | Unexplained "likelihood to enrol" score | Documented criteria; no automated offer decision |
| Fake intimacy | A message pretending to be a personal note | Clear sender, disclosure that an assistant wrote it, easy route to a person |
Never present an automated message as written by a named adviser who did not write it. And do not let the agent decide who receives an offer: it informs and follows up while the decision stays with your admissions team. Our piece on the role and limits of an admissions AI agent sets out the division of labour. If you are a registered provider, your regulator, TEQSA, expects accurate information for prospective students, which is one more reason to keep answers tied to an approved source.
Measure honestly
There is no reliable public conversion benchmark that fits every institution, so avoid borrowing one. Compare your own starting point with the position after launch, stage by stage: enquiries that became conversations, incomplete applications that were completed, unsubscribes, complaints. Treat unsubscribes as a quality signal. For the wider strategy, see the digital marketing guide for higher education.



