One intake door.
One coordinated case.
Human authority, always.
A multi-tenant operating system for First Nations prevention and family services — one universal intake, coordinated cases across every program, and AI that prepares the work while accountable people decide.
ZAVETBuilt for First Nations governments, Tribal Councils, and Indigenous service organizations.
A doctrine, not just a database.
Six principles shape every screen in ZAVET — so coordination, accountability and sovereignty are structural, not optional.
A person is not a case
Identity is sovereign and persistent. People exist in the system independent of any file opened about them.
An intake is not yet a case
Every concern enters through one door and is screened, triaged and routed before it ever becomes work.
A program file extends a case
Jordan's Principle, PMSS and every other program attach to the case — they never fragment the person's story.
Every deadline has an owner
SLAs, escalations and reminders are always assigned to a named person, never to a queue in the abstract.
Finalized records are amended
ZAVET is designed to preserve finalized records and record corrections as amendments.
AI prepares; humans decide
Automation drafts, summarizes and stages the work. A named, accountable person always makes the decision.
Every mandate, one coordinated case.
Purpose-built program workspaces designed around shared person records, coordinated cases and connected audit history.
Everything a coordinated service needs — in one system.
From the first referral to the final funder report, ZAVET replaces a patchwork of tools with one accountable workflow.
Universal intake & routing
One front door for every referral, screened and routed to the right program in minutes.
Coordinated case management
A single case spans programs, workers and time — no duplicate files, no lost context.
Work Queue with SLA & escalation
Every task carries a due date, an owner and an escalation path that never sleeps.
Finance & service authorization
Requests, approvals, budgets and payments tied directly to the case they serve.
Reporting with evidence maps
Reporting tools are designed to link calculated values to their source records.
Tier-3 AI report generation
Draft narrative reports and assemble evidence — staged for a named approver.
Multi-tenant data sovereignty
Each nation's data is designed to be isolated per tenant, with region and governance rules set for that deployment.
Mobile & offline field work
Home visits and field notes captured on-device, syncing when connectivity returns.
Full audit & provenance
Who did what, when and why — an audit trail designed to sit behind every recorded decision.
AI that prepares.
People who decide.
Three configurable tiers put generative AI to work on the busywork — drafting, summarizing and assembling — while the authority to decide stays where it belongs.
No AI action is final by design. Every output is staged for a named person to review, edit and authorize, and the platform is designed so that tenant data does not train third-party models.
Summaries, drafting help and smart search — every output human-reviewed.
Guided workflows, suggested next steps and pre-filled forms, always in context.
Report generation and evidence assembly, staged for a named approver.
Trust that is built in, not bolted on.
ZAVET treats data sovereignty as a first principle — designed into the architecture, and verified for each deployment before activation.
Canadian region deployment
Built to run in a Canadian region. Data residency is a deployment setting, confirmed for each tenant before activation.
Envelope encryption
Sensitive fields and secrets are designed for envelope encryption under externally managed keys. Key-management configuration is verified for each deployment before activation.
Row-level tenant isolation
Tenant isolation is designed to be enforced at the row level in the data layer. It is verified against each deployment's database configuration before activation.
Break-glass access control
Emergency access is designed to be time-boxed, scoped and recorded. Record retention and tamper-evidence are set per deployment.
Accessibility
WCAG 2.2 AA is the design target. Automated accessibility checks and keyboard operation are exercised in our end-to-end test suite. Colour contrast is not fully covered by those checks, and no conformance audit has been carried out.
No third-party model training
Tenant data is designed not to be sent to external or shared model training. Model routing and retention are fixed per deployment agreement and confirmed before activation.
Sign in to ZAVET
Sign in with your organization account. Access is designed to be role-based and actions to be recorded, and both are confirmed for each deployment before activation — data here is your Nation's, governed by your policies.