Documentation Index

Fetch the complete documentation index at: https://docs.aifabrix.ai/llms.txt

Use this file to discover all available pages before exploring further.

Evaluation guide

Prev Next

How enterprises evaluate AI Fabrix as a platform — before or alongside your first Azure deployment — using real identity, governed data access, and architectural fidelity (not feature demos alone).

Who this is for

Audience Use this section to…
Architect Confirm the operating model and trust boundaries fit your reference architecture
Security / compliance Validate identity, audit, and data-boundary behavior
Procurement / sponsor Assess structural suitability, cost predictability, and exit options
Platform team Run a time-boxed proof of concept with no throwaway work

Evaluation happens inside your Azure tenant with Entra ID (or your chosen identity path from tenant activation), real policies, and permission-aware dataplane access — not on synthetic demo tenants with service accounts.

How evaluation fits Foundation

Evaluate (this section)
  ↓
Marketplace install → Tenant activation → First session
  ↓
Pilot integration + certification (Build path)
  ↓
Scale (Adopt roadmap + platform sizing)
  1. Evaluation criteria and checklist — platform dimensions, security checklist, architecture review questions
  2. Proof-of-concept framework — 30-day PoC with exit criteria and risk/cost framing
  3. When approved: Azure Marketplace installationTenant activation wizard

What to validate (summary)

AI Fabrix should be judged on platform properties, not UI polish or pre-built agents:

  • Deployment and data stay in your Azure boundary
  • Identity preserved end-to-end (no anonymous AI service accounts for enterprise data)
  • Governance is structural — dimensions, protection, certification — not optional add-ons
  • Integrations expose governed capabilities (OpenAPI / Enterprise MCP), not raw credential handoff to models
  • Audit and evidence are deterministic, customer-owned
  • Lifecycle supports dev → test → prod promotion without re-architecture

Secondary topics (prompt libraries, model benchmarks, demo datasets) are useful but not primary pass/fail criteria.

Limits

This index summarizes evaluation criteria drawn from legacy evaluation material and architecture anchors. Specific Azure Marketplace SKUs, regional availability, and certification tooling details may change — confirm against your tenant activation output and current product release notes before procurement sign-off.

Related