ChatGPT
Claude
Perplexity
Gemini
GrokWhat AI qualification does for an Irish admissions office
An AI agent qualifies an enquiry by asking, in the chat, what an admissions officer would ask: which course, what prior qualification, which intake, how it will be funded. It records the answers, works out a priority and hands your team a readable record. It does not offer or refuse a place.
The job is narrow: decide who your team contacts first and with what context. Enquiries arrive at night, at weekends, around CAO deadlines and the Change of Mind window, and the evening before an open day. Without triage they wait in one queue.
This article covers the criteria to collect, an auditable score, handoff rules and the Irish safeguards: GDPR and the Data Protection Commission, the ePrivacy rules and the EU AI Act. For foundations, read our AI chatbot student recruitment guide. For a school-type view, see AI lead qualification for business schools.
A note on method: paid search volume and difficulty data were not available. This piece rests on public research and official sources and contains no performance statistics.
Criteria to collect: fit, intent, feasibility, reachability
Use the questions your officers already ask on the phone, and keep only those that change what happens next.
| Family | Example agent questions | Why it matters |
|---|---|---|
| Fit | Course of interest, prior qualification (Leaving Certificate, PLC, degree), undergraduate or postgraduate | Avoid calling someone who matches no course |
| Intent | Target intake, stage of the decision, comparing other colleges, request for a call or open day | Spot the applicant deciding in the coming weeks |
| Feasibility | Funding (family, SUSI grant, scholarships, employer), EU or non-EU status, documents on hand | Surface blockers before the first conversation |
| Reachability | Preferred channel, time window, consent to be contacted | Respect the person's choices and make the call-back work |
Three rules prevent drift. Every question must support a decision. Optional questions stay optional. Never rank people on sensitive characteristics such as ethnicity, religion or disability.
The Irish context shapes the questions
School leavers think in Leaving Certificate points, and undergraduate applications run through the CAO, while the colleges make the admission decisions. Postgraduate, mature and many international applicants apply directly to the college. The agent should explain how the process works, quote only the points the college publishes with their year, and never forecast whether a given applicant will get a place. Grants are handled by SUSI, and the agent should point there rather than estimate entitlement.
A two-axis score the team can audit
A useful score has two readable axes, fit and intent, instead of one opaque number. For every record your team should be able to say why the applicant is at the top of the list.
The example is illustrative and must be calibrated on your own data. The weights are not market benchmarks.
| Signal | Axis | Points (illustrative) |
|---|---|---|
| Specific course named | Fit | +2 |
| Prior qualification matches the entry route | Fit | +2 |
| Target intake in the current cycle | Intent | +3 |
| Asked for a call or booked an open day | Intent | +3 |
| Asked about fees or grants | Intent | +1 |
| Agreed to contact by phone or WhatsApp | Reachability | +1 |
| No matching course | Fit | Route to information, no sales call |
The total places the person in an action tier, never in an admissions decision.
- Priority: fast call-back from an officer, with the conversation summary.
- Follow up: information sequence, invitation to an open day or webinar.
- Nurture: long-term interest, useful content and a planned reminder.
- Out of scope: a helpful answer and a pointer to a better-matched course.
Set the rules before launch
When a score drifts from reality, correct it. Ask officers to flag misranked records, read conversations weekly at first and adjust. Keep a log of rule changes, because an applicant, your DPO or a regulator may ask. Our guide to lead scoring for student recruitment covers calibration on the CRM side.
Handing off to the team: when, how and with what
Handoff should follow explicit triggers, and the officer should receive a record they can use without rereading the chat.
Triggers
Hand off immediately in three cases: the person asks for a human, describes a sensitive situation or complaint, or the agent has no reliable answer. Hand off with priority when the score reaches the top tier or a call is requested.
What the record contains
- The name and contact details the person chose to give.
- Course, level, intake and study mode.
- The score and the signals that explain it.
- The conversation summary and unanswered questions.
- The consent captured, its channel and its date.
A call-back target per tier, set by your college, completes the setup. See our guide to lead routing and SLAs in student admissions.
Skolbot works in this pattern. Its web agent answers from the college's own content, qualifies the prospect, sends the conversation and qualification to the CRM, and a human team takes over. WhatsApp or phone follow-up can run from the CRM record. Specific CRM connections are confirmed in a demo, not assumed here.
Irish and EU safeguards: GDPR, ePrivacy and the AI Act
Qualifying prospects means processing personal data and sometimes profiling. Four points need a written decision before launch.
Lawful basis and transparency. The GDPR applies directly in Ireland, supplemented by the Data Protection Act 2018, and the Data Protection Commission is the regulator. You need a lawful basis, clear privacy information, a retention period and a position on processors and transfers. Ask your DPO whether a data protection impact assessment is needed. Applicants under 18 may use the chat, especially in CAO season, so write the notice in plain language.
Marketing messages. A chat is not consent to marketing. The Irish ePrivacy Regulations sit alongside GDPR and govern marketing by email, text and phone. Record what the person agreed to, through which channel and when, name your college in every message and make withdrawal easy.
Automated decisions. Article 22 GDPR restricts decisions based solely on automated processing that significantly affect people. The design principle follows: the score orders the call-back list, a person decides anything touching admission, and nobody is screened out without human review.
EU AI Act. The AI Act requires transparency when people interact with an AI system, so tell applicants in the first message and show a route to a person. It also lists systems that determine access or admission to education as high risk. A score that only prioritises a call-back is a different use, but do not use it to restrict access to information or an application, and have your DPO or counsel assess your case, as obligations apply in stages.
QQI is the body responsible for quality and qualifications, so the agent should state only the recognition and NFQ levels your college publishes.
On infrastructure, Skolbot's platform enforces EU data residency server-side. The assessment you keep on file is still your own.
Measuring without fooling yourself
Measure outcomes your office already tracks, not conversation counts: time to first call-back after an overnight enquiry, the share of records handed off with course and intake filled in, attendance at booked open days, and how often officers correct the score. Baseline your own figures before launch. Another college's conversion rate says little about your funnel, and we publish no figure here without a verified source.
Where to start
Start with one route, such as postgraduate or international enquiries, with three to five qualifying questions. Load course, fee and deadline pages, run the agent in front of your officers and fix content before the public sees it.
Then widen. Our grid for choosing an AI agent for student recruitment helps with evaluation, and AI agents for student recruitment in higher education covers the wider funnel.
