Deployment & data residency

Where your data lives — and how it runs

For Northwind Global, AIDE is deployed as a sealed Docker container inside the company’s own Microsoft Azure tenant — the data stays put; only an approved aggregate ever moves. The same sealed application can run in AWS, Oracle Cloud, or Google Cloud private tenants instead — pick the cloud you already trust.

Northwind Global — Microsoft Azure tenantWest Europe (Netherlands)Everything inside this boundary stays in Northwind’s cloud account.

Connected sources

Slack · Google · GitHub · HubSpot · AI-vendor telemetry — read in-boundary

Sealed AIDE container

Azure Kubernetes Service — sealed container

In-tenant AI scoring

Azure OpenAI — in-tenant endpoint

Private storage

Azure Blob Storage (private) · Azure Database for PostgreSQL · per-tenant git repo

→ Raw content & individual data never leave. Only a founder-approved, anonymized aggregate is transmitted for benchmarking.

Microsoft Azure — deployment stack

Active for Northwind Global

Compute (sealed container)

Azure Kubernetes Service — sealed container

AI model endpoint (in-tenant)

Azure OpenAI — in-tenant endpoint

Object storage (private bucket)

Azure Blob Storage (private)

Database

Azure Database for PostgreSQL

Secrets / key management

Azure Key Vault (KMS)

Identity & access

Microsoft Entra ID

Same sealed application, same guarantees — deployed into whichever cloud Northwind already trusts. Switching providers changes only the managed-service names above.

How it runs

Four steps, all inside your boundary

Deploy a sealed container in your tenant

AIDE runs as a sealed Docker application inside Northwind Global’s own Microsoft Azure tenant (AKS) — not our infrastructure. You hold the cloud account; we hold no copy of your data.

Read only shared-access content, in-boundary

It reads content the team already shares — department channels, group drives, wikis — plus AI-vendor usage telemetry. No inboxes, no one-to-one messages.

Build the graph & score on in-tenant models

The knowledge graph, RAG and AI scoring all run on model endpoints inside your boundary (Azure OpenAI / Bedrock / Vertex / OCI GenAI) — nothing is sent to any outside AI provider.

Two outputs, both in your control

Your full internal report stays in-tenant. A standardized, anonymized aggregate is produced for benchmarking — and your team approves it before anything is ever shared. Nothing auto-transmits.

The guarantees

The measurement is lower-risk than the status quo

EncryptedIn transit (TLS) and at rest (cloud-native KMS)

In your tenantRuns in Northwind’s own Azure account, not ours

Tenant-isolatedPer-tenant database, bucket prefix & git repo

Founder-approved egressOnly an anonymized aggregate leaves — after review

Audit-loggedEvery access to tenant data is recorded

Governance-readyDPIA + works-council support before any content is read

Read the full data-governance model in the CIO briefing. Next in the tour → Dashboard.