ibl.ai is available as a one‑click deployment on the Microsoft Azure Marketplace where universities can launch a fully managed instance inside their own subscription. Once deployed, ibl.ai relies on Azure OpenAI Service for large‑language models, AKS (or Container Apps) for its microservices, and the wider Azure stack—identity, data, and monitoring—to deliver secure, FERPA‑compliant generative AI at campus scale.
Key Azure Building Blocks
Azure OpenAI Service – direct access to all of OpenAI's latest models; ibl.ai chooses the best model per query while Azure handles GPU capacity.
Azure AI Studio & Content Safety – fine‑tune or ground models on university data, and apply Microsoft safety filters before answers reach students.
Azure Kubernetes Service (AKS) – container host for ibl.ai APIs, orchestration engine, and background workers; scales automatically during finals season.
Azure SQL / Cosmos DB – relational or NoSQL store for user profiles, transcripts, and analytics. Isolation can be per‑schema or per‑database to satisfy strict data policies.
Azure Storage – durable object storage for lecture uploads, embeddings, and backups, partitioned by tenant folder or container.
Azure Virtual Network + Private Endpoints – traffic stays on Microsoft’s backbone; each tenant can run in its own VNet with subnet‑level segmentation.
Microsoft Entra ID (Azure AD) – SSO for students and faculty; role‑based access control maps to tenant IDs for least‑privilege data access.
Azure Monitor & Application Insights – unified logs, metrics, and distributed traces power dashboards and auto‑scaling triggers.
How ibl.ai Uses Azure Day‑to‑Day
1. User query arrives. An Application Gateway routes HTTPS traffic to AKS pods running the ibl.ai API.
2. Model selection. The orchestration layer calls Azure OpenAI, picking GPT‑4o for rich tutoring or GPT‑3.5 Turbo for quick FAQ queues.
3. Context enrichment. Course PDFs in Azure Storage are chunked, embedded via Azure AI Search, and injected into the prompt.
4. Response & telemetry. The answer returns in <1 s; tokens, latency, and cost stream to Azure Monitor. Role‑based logs are stamped with TenantID for audit.
Why Azure Matters to Universities
Enterprise‑grade compliance – Azure certifications (FERPA, HIPAA, FedRAMP High) and Private Link keep student data locked down.
Deep Microsoft ecosystem – native hooks into Teams, Outlook, and OneDrive streamline faculty workflows.
Elastic scale, predictable cost – AKS autoscaling and pay‑per‑token OpenAI pricing prevent budget surprises.
Granular identity & RBAC – Entra ID ties AI access to existing campus roles; conditional access policies add extra safeguards.
Innovation runway – as Microsoft releases new models or new AI Safety features, ibl.ai adopts them with a config toggle.
By pairing Azure’s managed LLM platform with Microsoft’s secure cloud services, ibl.ai lets universities launch real‑time, multimodal tutoring within their own Azure tenant—no GPU procurement, no data leaving campus control.
Learn more at ibl.ai