Anar Labs
Anar Comply
Know your AI gaps before the regulator asks.

Describe an AI system once. Comply scores it against five GCC and EU frameworks and gives you the gaps in the order to fix them.

ForAI offices, compliance officers and CISOs
DateNovember 2026
The problem
Five frameworks. 61 requirements. Each in its own words.
12

UAE AI Ethics Guidelines

12

SDAIA AI Ethics Principles

10

Qatar National AI Ethics Framework

15

Oman AI and Digital Economy Framework

12

EU AI Act

Requirements per framework in the Comply catalogue, November 2026.

No single view

Each framework needs its own check against each AI system.

No order

A long list of gaps does not tell the team what to fix first.

No record

Last quarter's answers sit in a spreadsheet nobody can compare.

Anar Comply
How it works
One questionnaire. A score for each framework.
01
You describe the system

Purpose, limits, oversight, fairness tests and data handling, as yes or no and short text.

→
02
Anar Comply

Checks each of the 61 requirements and weights each result by its severity.

→
03
Your results

A score and a level for each framework, every gap, and the order to fix them.

Comply scores the answers you give. It does not inspect the live system. It calls no model.

Anar Comply
A worked example
A citizen assistant with oversight but no fairness tests.

Answers given

Purpose, limits and oversight described. Officer named. Override and stop control in place. No fairness testing.

Weighted score

8.5

Non-Compliant. 60 gaps: 26 critical, 28 high, 6 medium.

Weighted score per framework, out of 100

UAE AI Ethics Guidelines
15.4
Oman AI and Digital Economy Framework
11.8
EU AI Act
9.7
SDAIA AI Ethics Principles
5.6
Qatar National AI Ethics Framework
0.0

Output of Comply's assessment engine on this profile, run 18 November 2026.

Anar Comply
Fix order
The fixes come ranked by severity and effort.
#RequirementFrameworkSeverity, effortWhat to do
1Technical documentation (Annex IV)EU AI ActCritical, lowDocument design, development and validation per Annex IV.
2AI system certificationQatarCritical, mediumSubmit the system to the designated certification body.
3Cross-cultural fairnessQatarCritical, mediumTest fairness across nationalities and languages.
4Data privacy law complianceQatarCritical, mediumKeep a processing register and assess proportionality.
5Human decision authorityQatarCritical, mediumMap critical decisions and keep the human in charge.

The first five fixes from the same run. Wording shortened from the framework catalogue.

Compare runs and exportCertificate record, not yet signed
Anar Comply
Status today
What works now, and what comes next.
CapabilityStatusNote
Assessment against 5 frameworksShipped61 requirements, scored from your answers.
Gaps, fix order, compare, exportShippedFilter gaps by framework and severity.
CertificatesPartialA record with a number. No signature or verify step yet.
Evidence from Anar GuardPartialCoded. It does not connect to Guard today.
Evidence store and alerts on driftRoadmapPlanned.
ISO 42001, NIST AI RMF, Arabic contentRoadmapPlanned.
Anar Comply
Roadmap
Next: from answers to evidence.
Now
  • Five GCC and EU frameworks
  • Weighted score and level
  • Gaps in fix order
  • Compare and export
Next Roadmap
  • Evidence from Anar Guard
  • Evidence store for each control
  • Signed certificates
Later Roadmap
  • ISO 42001 and NIST AI RMF
  • Arabic framework content
  • Alerts when results drift
Anar Comply
See your gaps in one afternoon.

A one-month pilot: we deploy Comply in your environment and assess your first AI systems with your team.

anarlabs.ai/contact

Anar Comply