An after-hours process answers what it can, records the rest and names who replies
A good after-hours student enquiry response process does three things in order: it acknowledges the prospect straight away, answers only what your published content supports, and puts everything else into a queue that a named person works through first thing. It does not pretend the office is open. It tells the prospect honestly when a human will reply.
Most admissions teams already have an out-of-hours plan, but it is usually an auto-reply that says "we will be in touch". That sentence promises nothing and gives the adviser nothing to work with the next morning. The process below replaces it with something an adviser can act on.
This piece is written for UK universities and colleges. It covers the window you need to cover, triage, the record, the handoff, and the data protection rules that apply. For the case on why speed matters at all, read our article on how response time affects enrolments.
Define the after-hours window before you design anything
The after-hours window is every hour your admissions team cannot reply live, and in the UK it is wider than "after five o'clock". Evenings and weekends are the obvious part. Bank holidays, university closure weeks, staff training days and the gap between an adviser leaving and the next one arriving all belong in it too.
International prospects stretch the window further. An applicant in a different time zone writes at what is, for them, a sensible hour, and your office is closed. Treat time zone as a field you capture, so the morning adviser knows whether to email, call or wait.
| Window | Typical cause | What the prospect needs |
|---|---|---|
| Weekday evening | Prospect researches after school or work | A direct answer and a clear next step |
| Weekend | Families compare courses together | Answers plus an open day booking option |
| Bank holiday or closure | Office shut | An honest note on when the office reopens |
| Results and Clearing periods | Prospect holds a result and wants a decision | A fast route to a named person |
| Overseas time zone | Daytime for them, night for you | A reply at a time that suits them |
Write the window down. A process that lives in one adviser's head fails the first time that adviser is on leave.
Acknowledge at once, and say when a person will reply
The first message should confirm that the enquiry arrived, state that the prospect is talking to an AI agent, and give a realistic time for a human reply. "An adviser will reply by the end of the next working morning" is honest. "As soon as possible" is not a commitment.
Set the response target yourself. No public standard defines a correct reply time for a UK admissions enquiry, so choose a target your team can meet on a normal Monday and a bad one. A target you miss damages trust more than a slower target you hit.
The acknowledgement also protects you under consumer and advertising rules. The Office for Students expects providers to give prospective students accurate information, and anything an agent says on your behalf counts as information you published. Telling people when a human will respond is part of being accurate.
Answer only from published content
The agent should answer a question when your own pages contain the answer, and say "I don't know, here is who can help" when they do not. This is the single most important rule for out-of-hours cover, because nobody is watching the conversation in real time.
Good candidates for direct answers are:
- course content, structure and mode of study as published on course pages;
- published entry requirements and accepted qualifications;
- how and where to apply, including the UCAS route for undergraduate applications;
- open day dates and how to book;
- location, travel and accommodation information that is already on your site.
Do not let the agent interpret an individual's grades, promise an offer or confirm that a course has places. In the Clearing period, availability changes by the hour, so the agent should send the prospect to the live vacancy page and a person rather than state a status from memory.
Triage into three tiers
Triage sorts every after-hours enquiry by how fast and by whom it needs a human. Three tiers are enough for most teams.
Tier 1: routine questions the agent can close
These are questions answered fully by published content. The agent answers, offers an open day booking and logs the conversation. No human needs to act, although an adviser can still read it.
Tier 2: time-sensitive enquiries for a same-morning reply
A prospect with a result in hand, a query about an application that is about to be submitted, or someone asking to change course route is time-sensitive. The agent captures the details, tells the prospect a named adviser will reply first thing, and flags the record so it sits at the top of the morning queue.
Tier 3: sensitive matters that go to a named team
Disability adjustments, complaints, safeguarding concerns, mental health worries and anything involving a prospect under 18 should not wait in a general inbox. The agent should stop trying to help, give the prospect the right contact, and route the conversation to the named team responsible. If your institution has an out-of-hours emergency contact for students in crisis, the agent should be able to show it. Decide this with your safeguarding lead before launch, not after the first incident.
Capture a record the morning adviser can use
The record is what turns a night-time chat into a morning conversation. If the adviser has to ask the prospect to repeat everything, the cover has failed.
Capture a small, consistent set of fields:
- course or subject of interest and the entry route (for example A-levels, BTEC, Access course, international qualifications);
- contact preference: email, phone, or a call at a stated time;
- time zone, for overseas prospects;
- consent status, kept separate for answering the enquiry and for marketing;
- a short summary of the conversation and the questions the agent could not answer;
- the tier assigned and the time the enquiry arrived.
Prefer offering a callback to sending a cold message. A prospect who says "ring me after four" has given you a time and a reason to be ready. Our article on automating recruitment while keeping the human touch goes further into where the line sits.
Build the morning handoff queue and the rota
The handoff queue is a sorted list, owned by one person each morning, with the tier 3 items first, then tier 2, then tier 1 reviews. Ownership matters more than tooling. If "the team" owns the queue, nobody does.
Set a simple rota:
- a named owner for each working morning, with a named deputy;
- a separate owner for the first working morning after a weekend or bank holiday, because the queue is longer;
- extended cover during results and Clearing periods, agreed in advance with the people who will staff it;
- a written rule for what happens if the owner is absent.
Escalation should be explicit. If a tier 2 enquiry has not had a human reply by the target you set, the queue should surface it to a team lead. Without that rule, the oldest enquiries sink to the bottom.
The point of all this is to protect advisers' time, not to add to it. The agent clears routine questions so the morning goes to real conversations. Our piece on the time admissions teams spend on repetitive questions explains why that matters.
Data protection and marketing consent in the UK
An after-hours agent collects personal data, so UK data protection law applies from the first message. The Information Commissioner's Office explains the rules on direct marketing and electronic communications, which is the page to read before you let any follow-up message go out.
Four practical points follow:
- Tell prospects at the start that they are speaking to an AI agent and what you will do with their details. The EU AI Act sets a similar expectation in Article 50, and it matters if you recruit in the EU or buy from a provider that sells there.
- Answering an enquiry and sending marketing are different purposes. A person who asked about fees has not agreed to a newsletter, and PECR governs electronic marketing messages.
- Set a retention period for conversations and records, and apply it.
- Take extra care with under-18s, who make up a large part of early course research, and agree with your data protection officer what the agent collects from them.
Measure the process with your own data
Measure the steps you control, using your own records, and avoid comparing against numbers from elsewhere. Useful measures are:
| Measure | What it tells you |
|---|---|
| Time from enquiry to first human reply | Whether the handoff works |
| Share of handoffs answered within your own target | Whether the target is realistic |
| After-hours threads still unresolved after the first human reply | Where answers are missing from your site |
| Questions the agent could not answer | Content gaps to fix |
| Open day bookings made out of hours | Whether the booking route is used |
Review these weekly during recruitment peaks and monthly otherwise. Unanswered questions are the most valuable list you will get, because each one is a page you have not yet written. For the wider view on how AI agents and advisers divide the work, see our comparison of an AI chatbot and a human agent.
Where Skolbot fits
Skolbot deploys a website agent that answers from the institution's own content, qualifies the enquiry, books open days and appointments, and passes the conversation into the institution's CRM before handing it over to the admissions team. The school sets the response target and decides who owns the morning queue.



