Build my prototype

Schools · Colleges · Universities

Education intelligence

Institutions run on cycles that punish late decisions. An admissions shortfall discovered in July cannot be fixed in July. A student who disengages in week four is usually identified in week fourteen. The useful work is almost always earlier than the reporting allows.

  • 3 workflows
  • Pilot in weeks
  • No customers yet

Why this needs domain knowledge

A general-purpose model gets this wrong confidently.

Education data is seasonal, cyclical and full of cohort effects that generic forecasting handles badly. A model that does not understand intake cycles, term structure, or the difference between an applicant, an offer-holder and an enrolment will produce confident nonsense. The regulatory frame matters too: student data in India sits under the DPDP Act, and a meaningful amount of it belongs to minors.

Admissions intelligence Cycle to date
Enquiries 1,240
Applications 860
Offers 610
Enrolled 402

Forecast Enrolment tracking 6% under last cycle. Counsellor follow-up prioritised for 118 offer-holders.

Illustrative example · Education — sample data, not customer data.

The work

Three workflows, each followed all the way to an action.

Signal, decision, action. If a workflow stops at the decision it is not finished, and we have tried not to list any that do.

01

Admissions and intake forecasting

  • Application portal
  • CRM / enquiry system
  • Prior-cycle intake data

Signal

Application velocity, offer acceptance rates and enquiry-to-application conversion tracked against the same point in prior cycles, by programme and by source.

Decision

Where the intake shortfall is likely to land, which programmes are at risk, and which lever — outreach, scholarship allocation, counsellor time — is worth pulling now rather than next month.

Action

Counsellor follow-up prioritised, at-risk programmes flagged to the admissions head, and scholarship or outreach budget reallocated on approval.

02

Student at-risk early warning

  • Student information system
  • LMS
  • Attendance records
  • Assessment data

Signal

Attendance, assessment performance, submission timing and engagement trend together rather than as separate reports, at the individual level.

Decision

Which students are on a trajectory toward failure or withdrawal, how early the intervention window opens, and which type of support has historically changed the outcome for that pattern.

Action

The right faculty member or counsellor is notified with the specific reason, the intervention is logged, and the outcome is tracked so the model learns what actually worked.

03

Faculty load, placement and accreditation reporting

  • Timetabling
  • HR / faculty records
  • Placement tracking
  • Outcome data

Signal

Teaching load distribution, section-level outcomes, placement conversion by programme, and the evidence accreditation bodies will ask for, assembled continuously rather than in a panic.

Decision

Where load is unbalanced, which programmes are underperforming on placement, and which accreditation evidence is missing while there is still time to generate it.

Action

Load rebalanced on approval, placement effort concentrated where conversion is weakest, and accreditation documentation drafted from source records rather than reconstructed.

What we read

Typical data sources

  • Student information system
  • Learning management system
  • Admissions / CRM
  • Attendance and assessment
  • Placement records
  • Finance and fee collection

Read access or scheduled exports are enough to begin. Production write access is a separate, later conversation and is never needed for the predict and decide stages.

Where the boundaries are

Student data under the DPDP Act

Student records are personal data, and a significant share belongs to minors, which carries additional obligations around consent and processing. We design for data minimisation, scoped access by role, and retention that can be defended, and we will not process categories of student data the engagement does not need.

The model advises, faculty decide

An at-risk flag is a prompt for a human conversation, not a judgement about a student. Nothing in the system grades, ranks or makes a determination about an individual's future, and the reasoning behind any flag is always visible to the person acting on it.

Human opportunity here

What the recovered hours are actually for.

Administrative load in institutions is enormous and almost entirely invisible in strategy conversations. Time recovered from attendance reconciliation, report assembly and accreditation paperwork is time that can go back into teaching, counselling and student contact. That is the redeployment we would want to measure in an education pilot.

This is a thesis about where capacity should go, not a measured result. We have no deployments in education or anywhere else yet.

The full Human Opportunity thesis

In detail: the Capacity framework for measuring what the freed hours became, and the Pipeline from graduate to professional.

Starting

What a education pilot looks like

Small enough that being wrong is cheap, and specific enough that being right is provable.

Typical starting point

One programme group or one campus, one question: either intake forecasting for the next cycle or at-risk identification for the current cohort.

What we need from you

Read access to two or three years of historical data for the chosen question, and one person who understands how the process really runs.

What success looks like

An agreed measure set before we start. For intake, forecast accuracy against actuals at a defined checkpoint. For at-risk, how many students were identified early enough for the intervention to be possible.

Healthcare intelligence

Start here

Bring us the problem.

A sentence about what is slow, manual or missing is enough. We come back with a view on the fastest useful thing to put in front of you, and an honest answer on whether we are the right people to build it.

Build my prototype See how we work