Prevention was episodic
Families reached specialists too late, specialists worked in isolation, and administrators saw reports after the situation had already changed.
A new prevention paradigm after AI
Prevention can finally become a living system.
For decades, prevention systems were limited by cost, paperwork and the lack of continuous feedback. AI changes the technical and financial boundary. Prevention AI Platform is built as a hierarchy of applications: support for families, AI workspaces for specialists, and aggregate dashboards for institutions.
Why this must exist
AI makes a public-health prevention architecture possible.
The goal is not a chatbot. The goal is a new operating layer for the prevention of risks among growing generations: grounded in public-health evidence, connected to specialist practice, and continuously learning from aggregate, privacy-safe signals.
Support can be continuous
Interviews, diaries, de-escalation and specialist preparation can happen between formal appointments, at a cost that was previously impossible.
A hierarchy of apps
Each application is an organ of one system: consumer support, specialist workflow, scientific analytics and territorial governance.
Scientific layer
Implemented foundation
Taxonomy, knowledge base and BigQuery research loop
The platform is not driven by generic prompting alone. It already includes a prevention taxonomy, a structured knowledge-base direction and an analytics contour where aggregate events can support research and evaluation.
System Topology
Architecture of the Prevention Ecosystem
A seamless flow of data, insights, and support across three critical layers. Each layer respects privacy and ensures ethical AI use.
Roadmap as the main story
From a live consumer product toward multi-level prevention infrastructure.
Level 1
Implemented
Teenology: mutual trust for parents and teens
Free relationship check-up, AI Companion (not therapy), Family Bridge with Privacy Shield, guided sprints (Pulse + Battery), anonymized answer sharing, local diary and offline cards. Built for communication on equal footing — not parental control or surveillance.
Level 2
Implemented prototype
Prevention.AI: the specialist workspace
Specialists receive an AI assistant for quick consultation, methodological expertise and document drafts. The current specialist interface is already available as the Prevention.AI bot for professionals.
School stack
Live + pilot-ready
Schools: educator bot, specialist terminal, Teenology for families
Free AI for classroom teachers, Prevention Terminal for school psychologists, Teenology as the family layer on equal footing — mutual trust, not surveillance. Aggregate dashboards for leadership are the next funding milestone.
Level 3
Next funding milestone
Specialist terminals and dashboards for administrators
The next stage is to finish workstations and dashboards for every administrative level: school director, district coordinator, regional ministry and national prevention administration. Sensitive case data stays local; upper levels receive only aggregate signals.
Family Layer
Teenology: Mutual trust for parents and teens
Built for communication on equal footing — not parental control or surveillance.
- Relationship Pulse: Check-ins that bridge communication gaps.
- Privacy Shield: Teens control what they share.
- Offline Cards: De-escalation techniques available anytime.
Specialist Workspace
Prevention-AI for Public Schools & Clinics
Specialists receive an AI assistant for quick consultation, methodological expertise and document drafts.
- Case Management: Secure, compliant patient records.
- AI Analysis: Automated risk identification from check-ins.
- Evidence-based: Grounded in proven psychological protocols.
Data Integration
Aggregate Dashboards
Privacy-safe telemetry for school administrators and clinic directors to understand macro trends without exposing individual identities.
Open Access
AI Academy: Capacity Building
Deploying AI in mental health requires a profound shift in how professionals interact with technology. Our academy ensures safe, ethical, and effective adoption.
Why grants matter
The missing piece is not the idea. It is the runway to connect the layers.
Cloud and startup support would let the project move from implemented parts to a live end-to-end demonstration: long-context AI, scalable inference, specialist account federation, terminals and aggregate dashboards.
Microsoft path
Azure/OpenAI credits help scale deep AI accompaniment and move the specialist data layer toward production-grade PostgreSQL infrastructure.
Google path
Gemini on Vertex AI for long-context family sessions and specialist workflows; BigQuery for anonymized pilot metrics; Google Antigravity for agentic engineering velocity as a solo founder ships federation, terminals, and safety automation.
Get the Prevention-AI Terminal
Available for Windows. Secure, local, and built for privacy.
Platform Founder & Architect
Built by an MHPSS expert and practicing psychologist who writes the code.
Roman Dubrovsky, PhD brings 25+ years of experience in youth prevention, public health policy, and social research.
Career spans from conducting field research on child protection for UNICEF and WHO, advising governments, to managing national school psychology systems. As a CBT practitioner, Roman recognized the critical imbalance between administrative burdens and therapeutic work, leading to the creation of Prevention AI.
ORCID ID: 0000-0001-9876-9798