How Lumii keeps pupils — and your school — safe.
Lumii is the insight layer for child safeguarding and wellbeing in schools — built around real time pupil voice, powered by Lumii's proprietary Large Mental Health Model (LMHM). Lumii decides what to flag. Schools decide what to do about it. This page is the architecture behind it: the engage/route posture, the alert model, the AI rules, the data boundaries, and the regulatory alignment that make it safe to deploy at scale.
Lumii decides what to flag. Schools decide what to do about it.
Lumii makes thousands of flagging decisions a day — what tier an interaction sits at, what reasoning to attach, who needs to see it, when, with what context. No human team could make those decisions at that scale or speed. That is the work Lumii is built for.
What happens after a flag reaches an adult is a human decision, made by trained staff inside the school, governed by the school's existing safeguarding policy. Lumii does not act on alerts. Lumii does not contact pupils, parents, or external agencies. Lumii does not predict harm or make safeguarding recommendations.
This separation is deliberate and structural. The platform is designed so that safeguarding decisions cannot bypass human judgement, and so that the people making decisions about a child's wellbeing are the same people who have always made those decisions — DSLs, pastoral leads, counsellors, Heads — now with better information.
If a regulator, parent, or inspector asks "who is in control of safeguarding decisions in your school?" — the answer is unchanged by introducing Lumii. The trained staff in your school are. Lumii makes their job easier; it does not replace it.
Aligned with the UAE's agentic AI direction.
On 23 April 2026, the UAE government announced that 50% of federal sectors, services and operations will run on agentic AI within two years — under the directive of President His Highness Sheikh Mohamed bin Zayed Al Nahyan, announced by His Highness Sheikh Mohammed bin Rashid Al Maktoum. The UAE is positioning itself as the first government globally to operate at this scale through autonomous systems.
Lumii is built on this trajectory — not adapted to it. The LMHM is an agentic system: it analyses, reasons, executes routing decisions, and improves over time. The architectural difference is where the final decision sits: with the school's wellbeing and safeguarding team, every time. Agentic underneath. Human-led at the point that matters.
Lumii is the first agentic AI safeguarding and wellbeing platform built for UAE schools.
What Lumii does with what it hears.
Lumii doesn't only surface and route. The line between supporting a pupil and handing over to a trained adult is the nature of the issue, not the alert level.
Engages with the pupil
For everyday issues — friendship arguments, a hard day, low mood, exam stress, falling out with a sibling — Lumii engages directly. Reflection, calming, working through what happened. Where helpful, practical scaffolding such as drafting what they want to say to a friend. Most of this lives in Green.
Routes when it should
The same friendship argument, if it surfaces something safeguarding-related, is routed to a trained adult — and Lumii steps back. Lumii supports where support is appropriate. It hands over the moment it isn't.
The decision on every flag that reaches an adult sits with a human, by design.
Four levels. Every flag reviewed by a person.
Lumii's alert model is designed to be proportionate — most interactions warrant no action; some warrant attention; a small number warrant immediate response. This distribution holds across thousands of pupils and millions of conversations, and is consistent with what schools experience using other forms of pupil voice.
Green
No concern — routine emotional expression, daily check-in.
No action requiredAmber
Low–moderate emotional concern for staff review.
Routed to moderators in a daily summaryPurple
Safeguarding concern — action required.
Immediate alert to authorised staffRed
High emotional distress — urgent review.
Immediate alert to authorised staffAll Amber, Purple, and Red alerts are reviewed by authorised safeguarding staff. Purple and Red alerts are mandatory safeguarding requirements, not recommendations. The school dashboard records every alert, every review, every action — providing a full audit trail suitable for inspection.
From signal to action — the full path.
Pupil interacts
A pupil engages with Lumii through chat, check-in, story, or activity.
Lumii analyses
The interaction is assessed against the four-level alert model.
Flag raised
If above Green, an alert appears on the dashboard for authorised staff.
Human review
Trained staff review the alert and decide on the right response.
Action recorded
The school's response — escalate, support, monitor, close — is logged with a full audit trail.
Two voices, one alert structure
Lumii captures pupil voice and staff observations through the same alert structure. Members of the school's wellbeing and safeguarding team can add notes to a pupil's record or to a specific Lumii conversation, and flag them Amber, Purple or Red. The school sees both inputs in one place, with one triage flow, and one audit trail.
How Lumii's AI is built to behave with pupils.
The AI behind every Lumii interaction operates inside a strict set of behavioural rules. They are the line between what AI is permitted to do in a school and what only a trained adult is permitted to do.
Stays inside its lane
- Powered by Lumii's proprietary Large Mental Health Model (LMHM) — purpose-built for safeguarding and wellbeing in school contexts, trained on 40+ years of peer-reviewed research, operating in a walled-garden environment
- Surfaces patterns and routes flags — never decides outcomes
- Engages, reflects, and supports at Green and Amber — hands over to a trained adult the moment an issue needs one
- Steps back from anything safeguarding-related — that decision is human
- Does not diagnose, offer clinical advice, or predict harm
- Does not promise confidentiality to a pupil — every flag is reviewed
Built for pupils
- Calm, steady, non-clinical tone
- Age-appropriate content boundaries for the pupil's tier
- Culturally sensitive — no ideological, religious, or politically loaded language
- Encourages pupils to talk to a trusted adult
- Allows the pupil to disengage at any moment
The school knows who. Lumii doesn't.
Two architectural commitments sit underneath everything Lumii does. The first is about identity. The second is about data.
Identity stays with the school
The mapping between a unique code and a named pupil is held by the school, not by Lumii. Lumii receives a code; the school knows who that code belongs to. This is structural, not a policy choice — Lumii cannot identify an individual pupil even if asked to.
Data stays in the UAE, and isn't used to train AI
All pupil data is hosted in the UAE. We don't transfer it outside the UAE in normal operation, and pupil conversations are never used to train AI models. Staff wellbeing data is held in a separate system and is never visible to school leadership at the individual level.
Individual unique codes. Nothing else.
Each pupil receives a unique code from their school. There are no accounts to create, no passwords to remember, no QR codes, no personal identifiers required at any point. The school holds the link between a code and a named pupil. Lumii does not.
This matters because it means there is no pupil database for Lumii to lose, no account credentials to compromise, and no path from a Lumii alert back to a named child without going through the school. The school remains the holder of pupil identity at all times.
Built to fit the frameworks UAE schools work within.
Lumii is designed for the UAE regulatory environment from the architecture up — not retrofitted afterwards. The frameworks below are not aspirational; they shape every product decision.
Education frameworks
- ADEK SPIRE Wellbeing Mark — five dimensions (Self-empowerment, Physical, Intellectual, Relational, Emotional) with 3 elements each. Rolling out across Abu Dhabi private and charter schools from the 2025/26 academic year. Lumii maps to four of the five.
- ADEK Irtiqa'a — structured wellbeing data and audit trails for inspection
- KHDA / DSIB (Dubai)
- SPEA (Sharjah)
- APEA (Ajman, alongside MoE)
- RAKDOK (Ras Al Khaimah, alongside MoE)
- UAE MoE Safe and Responsible Use of Generative AI in Classrooms (2026)
- UAE National Education Charter — launched on Emirati Day for Education by HH Sheikh Mohamed bin Zayed Al Nahyan; defines the unified national vision for education across all schools. Lumii's pupil-facing platform is designed in alignment with the Charter's pillars on national identity and on physical and mental wellbeing.
- Emirati National Identity Strategy — established by the UAE Government to define and integrate Emirati identity across institutional and societal life, including educational curricula. Lumii's Emirati cultural model is anchored in this framework.
Federal & policy frameworks
- Federal Decree-Law No. 26 of 2025 — Child Digital Safety Law; enforcement Jan 2027; Lumii's architecture aligned with CDS Law expectations
- Federal Decree-Law No. 45 of 2021 — Personal Data Protection Law (PDPL)
- Wadeema's Law — Federal Law No. 3 of 2016; child protection duties
- ECA "Our First Five Years" strategy — aligned with the four pillars: health & nutrition, early care & education, family support, and child protection. Lumii contributes to the latter two.
Every concern, in one record, with the audit trail underneath.
Safeguarding architecture without case management is half a system. Lumii's case management layer closes the loop — every flag, every observation, every action and every closure held against the pupil's record, in one continuous trail.
Two sources, one record
A concern can be raised from pupil voice (a Lumii flag) or from a staff observation (the corridor conversation, the change a TA spotted, the parent who phoned in). Both flow through the same Amber, Purple or Red triage. Both land in the same case file. The DSL sees one record per pupil, not three half-records across three tools.
Every action captured
When the DSL opens a case, the observations, the relevant Lumii conversations and the teacher notes are already in view. Each action taken is timestamped against the staff member who took it. The case is closed with the reasoning written down — not lost in someone's memory three weeks later. Included in Lumii Core — not a separate module.
Why this matters in the UAE specifically. The UK has a connected safeguarding pathway — schools, social services and police visible to each other through shared systems. The UAE doesn't. Schools here operate within their own safeguarding policy and the reporting flows set by their regulator — which means the full record has to live inside the school, with the audit trail to evidence it. Lumii's case management is designed around that reality from the architecture up. See the operational detail on the Platform page →
Five questions to ask any wellbeing platform.
Useful as a procurement checklist when comparing approaches. The answers tell you a lot about the architecture underneath.
Does it hear from every child every day, or only when they choose to fill in a form?
Does every flag arrive with the reasoning behind it, or just a score?
Can your DSL pull a complete inspection-ready audit trail in one click, or does it get assembled at year end?
When safeguarding language appears, does the system route it to your authorised adults within seconds — or does it wait for someone to log in?
Who decides what happens next — the platform, or a trained adult who has read the evidence?
Lumii is built so the answer to the last question is always — and only — the trained adult.
Designed for cultural adaptability, not cultural assumption.
Lumii operates in schools across diverse cultural contexts. The platform's tone, language, and content boundaries are designed for adaptability — not for a Western default that gets translated.
- Lumii's cultural sensitivity is architectural, not aesthetic — purpose-built UAE International and Emirati cultural models anchor every conversation in the cultural reality of the child
- Aligned with the UAE National Education Charter and the Emirati National Identity Strategy — Arabic language, Islamic values, family centrality, national identity and belonging
- Available in 30+ languages including a UAE Islamic version — the reasoning happens in the language the child chose, not translated around an English core
- Neutral, dignified tone throughout — no slang, no humour at pupils' expense, no culturally inappropriate topics
- Avoidance of ideological content of any kind
- Respects family structures, faith traditions, and community values
- Designed for dignity and discretion in conservative cultural contexts
- Content boundaries reviewed regularly by clinical and cultural advisors
Real clinicians, governing real product decisions.
Lumii's clinical architecture is overseen by Dr Yebeltal Gezahegn, Consultant Neuropsychiatrist. Dr Gezahegn authors clinical content, reviews the safeguarding architecture, and ensures the AI rules remain consistent with current best practice in adolescent and child mental health.
This is not advisory in name only. Clinical review is a structural part of how the platform changes — alert thresholds, content boundaries, language patterns, and the behaviour of every model on the developmental ladder are all clinically governed.
What clinical oversight means in practice
- Every AI rule reviewed clinically before going live
- Alert thresholds calibrated against clinical safeguarding standards
- Content boundaries reviewed for age-appropriateness across each tier
- Re-entry guides and pastoral content authored or reviewed by Dr Gezahegn
- Neurodiversity traits model authored and clinically governed by Dr Gezahegn
- Regular review of platform behaviour against current best practice
Neurodiversity traits — clinical governance
The neurodiversity traits model is authored and clinically governed by Dr Yebeltal Gezahegn, Consultant Neuropsychiatrist. It is built on the same non-diagnostic principle that governs every other part of the platform — patterns and signals only, never diagnosis or labels. For a full description of what the model identifies, see the Platform page.
Safeguarding architecture, not safeguarding theatre.
A two-week trial. Setup in 1 hour. The same architecture you've just read about.