What's built, what's building, and what comes next.

Veille is at MVP stage, ahead of its first pilots. This page separates three things most roadmaps blur together: what is shipped, what is in active development, and what the founding cohort will steer.

StageMVP, Pre-pilot

Tests3183 passing

Roadmap shaped byDesign partner founding cohort

Available now, MVP What you can use today

AI system registry

Register, classify, and manage your AI system inventory. Risk tier, data categories, framework assignment, owner, and documentation links.

Guided intake form

Structured interview that determines applicable frameworks, risk tier, and initial documentation requirements for each system.

Article-by-article compliance evaluation

Maps each registered system against its applicable obligations, article by article, with gap tracing and evidence requirements.

AI-drafted compliance documents, bilingual, gap-triggered, gated by design

When a required document is missing, Veille drafts it in French and English, hashed immediately, gated for human review before it can close a gap.

Evidence vault, WORM + SHA-256 hash chain

Every artifact, uploaded or Veille-drafted, is hashed, timestamped, and stored with a tamper-evident chain.

Regulatory Watch, 6 official sources

Monitoring of 6 official Canadian and EU regulatory sources every 3 hours (Gazette officielle du Québec, Canada Gazette, OSFI, EU AI Office, EUR-Lex L series, Ontario newsroom). Material changes flagged in under 4 hours.

Bilingual audit dossier export

One-click export of all compliance artifacts for a given system, organized by framework, in French and English.

REST API + 15 frameworks (146 obligations)

Full API for programmatic access. Law 25 (23), PIPEDA (13), BC PIPA (12), AB PIPA (11), PHIPA (8), AMF Québec (8), OSFI E-23 (15), OSFI B-13 (8), CSA/CIRO (12), EU AI Act (10), TBS DADM (8), Bill 149 (1), Bill 194 (3), ISED Code (8), Human Rights AI (6), each with penalty, deadline, and evidence requirements.

Building, Q3 2026 In active development

Automated discovery connectors, MLflow, Databricks, GitHub, SageMaker Available (beta)

Read-only connectors that scan your ML infrastructure for unregistered AI endpoints. MLflow, Databricks, GitHub, and SageMaker connectors are built, unit-tested, and wired into the credential-driven discovery loop, now in beta pending validation against live customer infrastructure. No write access, metadata only. Connector credentials encrypted at rest.

Live compliance scan, validated by external reviewer

First live scan against a real AI system (not simulated), with output reviewed by a qualified Canadian privacy/AI law firm for accuracy validation.

Legal reviewer partnership, cabinet en formalisation

Formal engagement with a qualified Canadian AI/privacy law firm for review of material regulatory changes before encoding. In commercial discussion.

SLA structure, Tier 1 seven-day response

Formal SLA with service credits for Tier 1 framework changes. Detection <4h, re-evaluation within 7 days of a material change.

Planned, Q4 2026 Next on the horizon

Azure Machine Learning connector

Read-only scan of Azure ML workspaces for registered models and production endpoints.

SOC 2 Type II audit initiation

Engagement with a qualified auditor; the observation period begins as the first paying clients come on board, with the report targeted for Q4 2026.

Multi-tenant organization model

Support for clients managing compliance across multiple subsidiaries or business units under a single account.

AIDA successor framework

AIDA (Bill C-27) died at prorogation in January 2025. No federal AI bill has replaced it since. Bill C-36, tabled 15 June 2026, is a privacy reform that creates automated-decision duties. Veille will encode the applicable obligations once they are enacted.

What's built, what's building, and what comes next.

What does your team need that isn't here?