AI the public can trust with public money
Personalize learning, route services and surface fraud on your own governed data — auditable by design, equitable by mandate, proven across 500K+ learners and three education systems.
A Hominis app module · your public sector & education data, app-ified
Better outcomes the public can audit
Personalize learning, route services and surface fraud on sovereign, governed data.
Adaptive learning at population scale
Curriculum, pacing and assessment that adapt to every student's knowledge state — knowledge-tracing models running across whole districts, not pilot classrooms.
50% lower dropout · 28% higher performance
Service delivery & case triage
Citizen requests, benefits claims and permit applications routed, prioritized and pre-checked so caseworkers spend their hours on judgment, not queue-sorting.
Faster resolution, fewer backlogs
Fraud, waste & improper-payment detection
Anomaly models flag improper claims and procurement irregularities before money leaves the treasury — every alert traceable to the signals that raised it.
Recovered spend, audit-ready flags
Budget allocation & spend transparency
The BudgetSankey instrument traces every public euro from appropriation to outcome, so program officers and auditors see where funding flows and what it bought.
Full allocation-to-outcome traceability
Educator & caseworker copilots
Natural-language explanations and lesson plans that give teachers and frontline staff back hours each week — oversight kept firmly in human hands.
~5 hours/week saved per educator
Auditable by design, sovereign by default
What survives procurement, oversight and the public record.
Auditable decisions, not black boxes
Every recommendation — a learning path, a flagged claim, a budget line — carries the evidence trail that produced it. Attention-based explanations and the BudgetSankey instrument make each decision interrogable by auditors, ombudsmen and the public. In government, defensible reasoning, not raw accuracy, is what survives oversight.
Runs on sovereign, governed data
Models train and serve inside the institution's own data perimeter — student records, benefits data and treasury systems never leave public custody. Data residency and GDPR become structural properties of the deployment, not compliance bolt-ons, which is what clears the security review every public buyer runs.
One method, tuned for public sector & education
Assess, Transform, Sustain — the cycle every organization runs, dropped one level deeper for your sector's pains and sticky aspects.
Assess
Map siloed agency data against equity, FOIA and procurement constraints
We start where public programs actually stall: data scattered across an LMS, a benefits platform and a treasury ledger that were never meant to talk. The Assess phase inventories those sources, maps lawful-basis and data-residency boundaries for each, and pressure-tests candidate use cases against equity impact and public-records obligations — so the first model shipped is one that can survive an audit, not just a demo.
Transform
Pilot one high-stakes service in one cohort, with human oversight from day one
Transform means standing up a single accountable slice — adaptive learning for one district, claims triage for one benefit line — with educators or caseworkers kept in the loop and an override path on every recommendation. Models train on governed data, every output ships with its explanation, and success is measured on the outcome the public cares about, not internal accuracy metrics.
Sustain (AIOps)
Scale to population level while holding equity, drift and public-trust accountability
Sustain is where public-sector AI usually breaks: a pilot that worked for 2,000 students has to hold for 500,000 without quietly disadvantaging a subgroup. We monitor for subgroup drift and disparate impact, keep BudgetSankey traces and decision logs live for oversight bodies, and version models so any auditor can reconstruct why a citizen got the answer they did.
Learning paths that adapt to every student
Adaptive curriculum delivery — student performance up 28%, engagement up 40%.