How much does your school really know about every prospective student — and the family behind them?
Admissions now spans every level of education — a parent choosing a school, a student comparing master’s programs, MBAs, and EMBAs — and at every level the application is preceded by the same long trail: dozens of website visits, brochure downloads, open days, campus tours, chat threads, phone calls, meetings. And yet in most admissions systems, all of that history fits into five fields: name, phone, intended program and entry, source, and follow-up status.
This is not a discipline problem, and more training will not fix it. The gap is structural: a CRM manages the information counselors are asked to enter, while the enrollment process keeps producing signals no one has time to type in. The next generation of admissions systems starts from a different premise — manage the signals, not the entry. We call this direction Enrollment Intelligence.
CRM’s blind spot is by design
Consider a hypothetical prospect — a composite of patterns every admissions office will recognize, at every level from K-12 to higher education. Over six weeks, one applicant visits the school website seventeen times, returns to a program page six times, downloads the admissions brochure twice, attends an open day, exchanges twenty-three messages with a counselor on the school’s enterprise messaging platform, and calls three times — the last conversation about tuition and financial aid.
What does the CRM show? Name: Zhang. Intended entry: Fall 2027, a master’s program. Source: Open Day. Status: in follow-up.
Everything else lives in a counselor’s memory, a personal notebook, chat histories, and a private spreadsheet. When that counselor is unavailable, the context leaves with them. When an applicant quietly goes quiet, the system has nothing to say.
The instinctive fix — asking counselors to log more — has been tried for a decade, and it fails for a structural reason: every extra field is one more step in an already overloaded day. Counselors, like teachers, are time-scarce professionals who set aside anything that adds net burden; a tool that demands more entry loses to the phone call waiting in the queue (see The Software Is Live. Why Aren’t Teachers Using It?).
If the loss is built into the design, the fix has to be architectural: capture at the source, so that understanding no longer depends on anyone’s typing.
Every interaction is a signal
An enrollment process produces three families of signals:
- Digital: website visits, page views, brochure downloads, form submissions, email interactions;
- Human: messaging conversations, phone calls, meetings, open days, campus visits;
- Operational: CRM records, spreadsheets, applications, assessments, offers, enrollment outcomes.
The opportunity is not to collect more data. Most schools already sit on more signal than they use. The opportunity is to connect what already exists — and the richest untapped sources are the conversations, because that is where intent, concerns, and objections actually get said. A page view can tell you that a prospect looked at the program; only the conversation tells you why they hesitated.
When conversations become data
This is where the current shift differs from a decade of CRM pitches. Speech-to-text and language models have made unstructured conversation — the majority of enrollment signal — machine-readable at a usable cost. A phone call becomes a transcript; a transcript becomes structure: the topics discussed, the questions asked, the needs expressed, the objections raised, the commitments made, and the natural next action.
The difference is easiest to see side by side. The follow-up note in a traditional system reads: “Zhang asked about the curriculum. Follow up.” An intelligence-grade record reflects what actually happened: which programs were discussed, what the applicant’s stated concerns were — fit, cost, outcomes — what was promised, and what should happen next, carried forward automatically instead of from memory.
One boundary has to be stated plainly, because in education it is not optional: conversation data can only flow where the school’s enterprise communication platform, APIs, and permissions allow it, and where recording, consent, and data-governance requirements are met. Where those conditions are not in place, this layer simply does not happen. What the technology makes possible, governance decides.
From understanding to the next best action
Understanding is only worth what it changes. The point of connecting signals is not a richer profile; it is a better decision about the next hour of a counselor’s day: which prospects to contact today, why them, what to talk about, what to send, and who on the team should own it.
This is also what changes a dashboard into decision support. Traditional reports count: leads, applications, offers, conversion. An intelligence view prioritizes: which high-intent prospects have not been followed up, whose engagement jumped this week, which applications are stalling, which offers have gone quiet — each with a recommended action attached.
Nor does the recommendation wait to be asked for. When an applicant’s status changes — engagement jumps, a new concern surfaces, an application stalls — AI can proactively push a follow-up suggestion to the responsible counselor, based on the applicant’s information and state: whom to contact, why now, and what to talk about.
AI can also do something a traditional CRM cannot: surface current students and alumni who resemble the applicant — similar academic backgrounds, interests, or paths through education. Compared with “our school fits you,” the real story of someone who walked a similar path is often the most persuasive information a counselor can bring to the conversation.
And judgment stays human. A recommended action is decision support, not a command; the counselor owns the relationship and the call. Schools that have adopted AI tools before know why this matters: usage follows trust, and trust follows control.
From system of record to Enrollment Intelligence
It helps to see where this sits in a longer arc. Stage one is the system of record: it records what happened. Stage two adds analytics: it understands what happened. Stage three is what we are discussing: it recommends what to do next.
That ladder, to be fair, is not new — the industry has drawn it for years. What is new is the enabler. As we argued recently, the pilot era for large language models is over; understanding unstructured text and speech, at the scale and cost admissions requires, has crossed from demo to production (see The AI Pilot Era Is Over: It’s Time for Education to Move Beyond Experimentation). The third stage stopped being a slide and became a buildable system.
For schools, this is Enrollment Intelligence: the admissions instance of what we call Education Intelligence — AI that works from the school’s own rubrics, data, and context, not from generic prompts. The future of admissions is not a better CRM. It is an intelligent system that understands every interaction, recognizes every signal, and helps schools act at the right moment.
If you are rethinking how your admissions team captures and uses the signals it already generates, we would love to hear from you — contact us.
