OpenIndustrial Documentation
Menu

Roadmap

Where Governed AI is heading. The Control Plane deepens its moat.

This page tracks where OpenIndustrial is heading. Each item reinforces the competitive moat: AI-agnostic, governance-first, de-risked entry.

Status Overview

FeatureStatusCategory Impact
MCP IntegrationSHIPPEDAI-agnostic foundation
Governance LoopSHIPPED"Nothing without approval"
Warm QueriesSHIPPEDAI-assisted query building
Fire AziSHIPPED"Fire us, keep running"
Schema PromotionIn developmentGovernance scales
Persistent MemoryPlannedGovernance learns
Autonomous ActionsPlannedGoverned automation
Proposal BucketsPlannedMulti-AI organization
Governed MCP GatewayPlanned"Bring any MCP" governed
Multi-Cloud Control PlaneVisionGovernance anywhere

What's SHIPPED (Start Here)

Thinking Tip:

Everything below is working today. Start with the Quickstart to see it in action.

MCP Integration

Status: SHIPPED

Connect any MCP-native AI to your Control Plane. Today: Claude natively, GPT/Copilot via adapters. Tomorrow: any AI that adopts the standard.

Get Started →

Governance Loop

Status: SHIPPED

Every action queues for your approval. Complete audit trail. Nothing happens without your explicit decision.

See How It Works →

Warm Queries with Azi

Status: SHIPPED

AI-assisted KQL query building. Natural language to structured queries. Saved, version-tracked, reusable.

Build Your First Query →

Fire Azi and Keep Running

Status: SHIPPED

Remove Azi anytime. Your data stays in your Azure tenant. Your queries keep running. Your configurations persist.

Understand Independence →

In Development

Schema Promotion

Status: In development | Expected: Q2 2026

What it means:

  • Observed data patterns become confirmed schemas
  • AI proposals can only act on promoted structure
  • Version history tracks every promotion decision

How it deepens the moat:

  • "Nothing without approval" extends to structure itself
  • Governance isn't just for actions - it's for meaning

Learn More →

Planned

Persistent Memory

Status: Planned | Expected: Q2 2026

The Control Plane will remember patterns from your approvals. Not autonomous learning - governed pattern recognition.

What it means:

  • System notices when you approve similar actions
  • Proposes reflexes based on confirmed patterns
  • You approve or reject the reflex itself

How it deepens the moat:

  • "Nothing without approval" - reflexes are approved patterns, not autonomous rules
  • Your approval history becomes valuable infrastructure

Autonomous Actions

Status: Planned | Expected: Q3 2026

Pre-approved action patterns that execute within boundaries you define.

What it means:

  • Define rules: "If X happens, do Y automatically"
  • Rules themselves require approval
  • Full audit trail of automated actions

How it deepens the moat:

  • "Nothing without approval" - even autonomous actions have approved boundaries
  • Governance, not autonomy, is the foundation

Learn More →

Proposal Buckets

Status: Planned | Expected: Q3 2026

Organize proposals from multiple AI sources into named buckets.

What it means:

  • Separate Claude proposals from Codex proposals
  • Scope buckets to users, teams, or surfaces
  • Manage multi-AI workflows

How it deepens the moat:

  • "Bring any AI" - practical multi-AI support

Learn More →

Governed MCP Gateway

Status: Planned | Expected: Q4 2026

The challenge: As enterprises adopt MCP (Model Context Protocol) for AI tool integration, they face a governance gap. MCP doesn't provide native frameworks for policy, authorization, and audit. Emerging MCP gateways (Obot, TrueFoundry, Kong) govern tool CALLS - but they're just a layer you add. They don't have the data. They don't have the platform.

What it means:

  • Governed Catalog - IT-verified, tested MCP servers with role-based visibility
  • Governed Access - Per-tool permissions, parameter-level policy, surface scoping
  • Governed Approval - MCP calls can generate proposals (same queue as Azi)
  • Context Optimization - Dynamic tool loading, schema optimization, warm query exposure
  • Full Observability - Every MCP call logged with identity, cost tracking, compliance-ready

How it deepens the moat:

  • "Nothing without approval" extends to MCP capabilities
  • No "Shadow MCP" - every capability is governed
  • Industrial + MCP governance is an unclaimed intersection
  • Same governance whether AI or human initiates
  • Control Plane, not Gateway - we govern 10 layers, not just calls

Learn More →

Vision

Multi-Cloud Control Plane

Status: Vision | Expected: 2027

The Governed AI Control Plane today runs in your Azure tenant. The architecture is ready for more.

The foundation is already here:

  • "Fire us and keep running" proves the portability foundation
  • Schemas, surfaces, and queries are already exportable
  • Governance logic is declarative and cloud-independent

What it means:

  • Deploy the Control Plane to AWS, GCP, or on-premises
  • Same governance, same approval workflows, same audit trails
  • Your investment today becomes your foundation everywhere

Why this matters:

  • 89% of enterprises operate in multi-cloud environments
  • Governance shouldn't be locked to a single cloud
  • "Nothing without approval" works the same on any infrastructure

Learn More →

Category Evolution Philosophy

Every roadmap item reinforces the competitive moat. We're not adding features - we're deepening the category.

ItemHow It Reinforces the Moat
Schema Promotion"Nothing without approval" extends to structure
Persistent MemoryYour approvals become valuable infrastructure
Autonomous ActionsEven automation is governed
Proposal Buckets"Bring any AI" at scale
Governed MCP GatewayControl Plane for AI capabilities - govern data, tools, calls, platform
Multi-Cloud Control PlaneGovernance travels anywhere - not locked to one cloud

Start with What's Shipped

Don't wait for roadmap items. The foundation is working today:

Thinking Tip:

The pilot is how we BOTH learn if this is right. $15K. 4 weeks. Your Azure tenant. Run an experiment, not make a bet.

On this page