Skip to main content
NewContriqx is now in private beta.Request access →
The platform

A control plane for AI that compounds.

Four layers — interface, capabilities, intelligence, foundation — connected by a conversational advisor that speaks your regulations and your data, in English and Arabic.

Architecture

Four layers. One coherent system.

From the interfaces your team uses, through the capabilities that do the work, down to the intelligence engine that ties it all together — and the security foundation underneath.

Layer 01

Interface

Where users interact
Conversational Advisor

Ask anything in English or Arabic. Spans all capabilities.

Web Application

Bilingual UI with RTL support. Role-based views.

REST API & MCP

Programmatic access. Integrate with your existing stack.

Integrations

GRC tools, ticketing, SSO, identity providers, cloud audit logs.

Layer 02

Capabilities

Where governance work happens
Inventory Engine

Every AI system, owner, vendor, risk tier — in one registry.

Workflow Engine

Approval chains across compliance, legal, risk, security, business.

Continuous Assessment

Re-evaluates whenever the model, vendor, or regulation changes.

Evidence Mapper

Maps obligations → controls → evidence with traceable chains.

Guardrails Engine

Policy-as-code enforcement at the model boundary. Arabic-aware.

Layer 03

Intelligence

Where intelligence compounds
Knowledge Graph

Entities, relationships, evidence chains. Reasons over your AI landscape — every system, risk, control, and regulation, connected.

Semantic Search

Regulatory embeddings and document retrieval. Grounds every answer in source material — not a public web search.

Arabic NLP Engine

Native Arabic entity extraction, evaluation, and guardrails. Not a translation layer — built from the ground up for Arabic content.

Layer 04

Foundation

Security, scale, trust
Dedicated tenancy

Your own instance. No shared infrastructure.

Enterprise identity

SSO · SAML · SCIM · RBAC.

Forensic audit log

Tamper-evident, append-only, exportable.

Deploy anywhere

Cloud · on-prem · air-gapped.

Data sovereignty

Residency in EU · KSA · UAE.

Encryption & keys

BYOK and customer-managed keys.

Every query into the Advisor traverses all four layers. Every document ingested updates the knowledge graph below. Every guardrail runs against the same source of truth. The platform compounds — it doesn't just store.

Spans all five capabilities

Ask anything. In English or Arabic. Get sourced answers.

The Contriqx Advisor is a conversational AI layer that sits across every capability. Ask "which approvals are stuck on Legal?" or "are we ready for SDAIA's high-risk classification?" and get sourced answers grounded in your own governance landscape — not a public web search.

InventoryWorkflowsAssessmentEvidenceGuardrails
C
Contriqx Advisor
Cross-capability query
Which AI systems are blocked from deployment this quarter, and why?

Three systems are currently blocked.

  • Customer Churn Predictor v2 — awaiting Legal sign-off · 3 days remaining · approval chain at step 3/4
  • Loan Pricing Engine — risk-adjusted ROI below threshold · 8 of 14 controls in place · pending stakeholder alignment
  • Branch Footfall Forecaster — Arabic bias evaluation overdue · 47 days past last assessment · SDAIA flag raised

Sourced from inventory · workflows · continuous assessment

Works with your stack

Not another silo. The layer on top.

Contriqx integrates with the GRC, audit, and data-governance tools you already run — then sits above them as one connected view. As your AI footprint grows, Contriqx stays the single place your board, regulators, and auditors look to answer one question: "is our AI governed?"

GRC platformsAudit & evidenceData governanceIdentity & SSOCloud & model providersTicketing
In action

Three scenes from a live Contriqx tenant.

AI governance fails when it lives in a security team's spreadsheet. Contriqx makes it the work of the whole business — with the visibility, workflow, and accountability to match.

01
Visibility · For everyone, not just engineers

Your CFO sees the same AI landscape as your CISO.

One AI system has a finance owner, a compliance owner, a security owner, and a model owner. Contriqx gives each of them the view they need — same source of truth, different lens. No exports. No reconciliation. No "let me get back to you."

contriqx.ai / systems / super-app-chat-agent
SA
Super App Chat Agent
Executive AI oversight snapshot
Compliance coverage
83%
↑ 5% vs Q1
Critical systems
4
2 at risk
Approval median
13 days
Evidence completeness
88%

Same system. Switch persona → different lens.

02
Pre-implementation · Decide before you spend

Catch a bad AI investment before it ships.

Most AI governance happens after deployment, when fixing things is expensive and political. Contriqx assesses business value, regulatory exposure, and operational risk at the proposal stage. Approve, reshape, or kill the project before a single model is trained.

contriqx.ai / proposals / loan-pricing
!
AI loan pricing engine
Proposal · EU AI Act: High-risk
Verdict
Pause. Close 8 controls before approval.
Required controls in place6 / 14
Stakeholder alignment3 / 5
Estimated ROI
€2.1M / yr
Risk-adjusted
€0.4M / yr
03
Workflow · Accountability without bottlenecks

Sign-offs in days, not quarters.

The DPIA, the security review, the legal sign-off, the executive approval — Contriqx threads them through one workflow with the business context attached. Each approver sees what's relevant to their lens, signs off in-app, and the evidence lands in the audit trail automatically.

contriqx.ai / sign-off / churn-v2
Churn Predictor v2 — release approval
Started 4 days ago · 12 evidence docs attached
In review
  1. Data Protection Officer
    Lena Hartmann
    2 days
  2. Information Security
    Tarek Al-Mansour
    1 day
  3. 3
    Legal Counsel
    Sophie Berger · in progress
    3 days left
  4. 4
    Business Owner
    Marco Rossi · awaiting Legal
    queued
Estimated close: 4 business daysNo chase emails sent
Arabic-native — not a translation layer

Same question. Native handling. In both languages.

Governing AI that processes Arabic content requires Arabic-aware extraction, evaluation, and guardrails. No other AI governance platform does this. We do.

English
Are we ready for SDAIA's high-risk classification?
Partial. 4 of 6 SDAIA high-risk obligations are met. Missing: model card publication, Arabic bias evaluation. Owner: Risk & Compliance.
العربية
هل نحن جاهزون لتصنيف سدايا عالي المخاطر؟
جزئيًا. ٤ من ٦ التزامات سدايا عالية المخاطر مستوفاة. المفقود: نشر بطاقة النموذج، تقييم التحيّز باللغة العربية. الجهة المسؤولة: المخاطر والامتثال.

Arabic NER · Arabic-aware guardrails · Arabic document ingestion · Arabic safety evaluation

Ready to see this on your own AI?

Share documentation about one AI system, one policy, or one workflow. We'll map it to your frameworks and send you a written report — in five business days, no sales theatre.

Get your free report →Back to home