Bloom Cap
An assessment and counselling platform helping students identify their true potential — hardened with the role and privacy model an education business actually needs.
What they needed — in their own words
“Students take assessments, counsellors review the results and run sessions, and parents see progress reports. A parent must never see another family's data, and counsellors' private session notes must stay private.”
The system they run on today
- Assessment engine with scoring and student reports
- Counsellor scheduling and session management
- Separate student, parent, and counsellor portals
- A visibility model enforced at the data layer, not the interface
Arrived, assessed, hardened
Every Neolen engagement follows the same path: an honest assessment of what's actually there, then the hardening work that makes it a system. Here is how it played out on this one.
What arrived
A working assessment product with three kinds of users and one quiet problem: access control lived in the interface. The screens hid the right things, but the database would hand any authenticated user any record they asked for.
What the assessment found
The assessment confirmed authorisation was never enforced at the data layer — the classic gap between 'login works' and 'access is controlled'. It also surfaced unanswered policy questions: what does a parent see when a student turns eighteen mid-programme? What about two guardians with separate accounts?
What we hardened and shipped
We enforced row-level access rules in the database itself, resolved the guardianship edge cases with Bloom Cap directly — report visibility follows guardianship, session notes remain counsellor-only at every age — and wrote the rules down where the next engineer can find them.
What the client runs on today
Bloom Cap operates the full assessment-to-counselling funnel with a visibility model that has held through every subsequent change, because it's enforced where the data lives rather than remembered in the interface.
Added to the playbook
Interface-only authorisation is now the first thing our audit harness checks on every multi-role application — it's the single most common critical finding in AI-built apps.
Hit a wall with something you've built?
The audit is fixed-fee and standalone — a plain-language report of what's actually wrong, ranked by risk, yours whether or not you continue.