Saudi GRC terms, defined
The authorities, frameworks, and concepts that come up in Saudi cybersecurity and data protection compliance — with the Arabic term alongside each, because most practitioners work in both.
This page describes how we approach this work and how Govrly supports it. Where regulatory requirements are mentioned, confirm what applies to your organisation against the source documents published by the relevant authority.
Regulators and authorities
A Saudi organisation rarely answers to just one. Knowing which body sets which requirement is the first scoping decision in any programme.
The national authority responsible for cybersecurity in Saudi Arabia. It issues the Essential Cybersecurity Controls and the specialised control documents that sit above them, and receives compliance submissions from in-scope entities.
The central bank and financial-sector regulator. It issues the Cyber Security Framework that applies to the financial institutions it supervises, assessed on a maturity model rather than on implementation status.
The regulator for the ICT, telecom, postal, and space sectors, formerly known as CITC. It issues the Cybersecurity Regulatory Framework that applies to its licensed and registered providers.
The authority responsible for data and artificial intelligence nationally, and the enforcing body for the Personal Data Protection Law. Breach notifications and controller registration go to SDAIA.
Operating under SDAIA, NDMO issues the national standards for data governance, classification, and management. Its policies are directed at government data specifically, and data residency authority sits here rather than in the ECC.
Sets digital government policy, including cloud adoption direction. Digital service design must account for SDAIA privacy requirements and NCA cybersecurity requirements, which makes service design a compliance decision from the outset.
Frameworks and laws
These overlap heavily at control level. What differs is scope, assessment method, and which authority reads the result.
The NCA's mandatory national baseline. The current edition, ECC-2:2024, has four main domains, 28 subdomains, 108 main controls and 92 subcontrols. It applies to government entities, their affiliates, and critical national infrastructure.
The superseded edition, with five domains, 32 subdomains and 114 controls. ECC-2:2024 restructured the industrial control systems domain, merged overlapping subdomains, and moved data residency authority to NDMO.
Issued in 2017 under Circular 381000091275. Four domains and 32 subdomains, assessed on a six-level maturity model from 0 to 5, with level 3 as the minimum expected. Business continuity is explicitly out of scope.
Issued by CST under Decision 424/1442. Six control domains and four pillars — strategy, people, technology, and processes — with three compliance levels assigned by the regulator. Aligns with ISO 27001 and NIST.
Issued by Royal Decree M/19 in 2021 and amended by M/148 in 2023. Enforced by SDAIA, which supervises how organisations collect, process, store and transfer the personal data of individuals in the Kingdom.
The international standard for information security management systems. Mandatory requirements sit in clauses 4 to 10; Annex A lists 93 controls across four themes. Voluntary and independently certifiable, unlike the Saudi frameworks.
The control list in ISO 27001:2022 — 93 controls in four themes: Organizational (37), People (8), Physical (14), and Technological (34). Eleven controls were new in the 2022 edition, including threat intelligence and secure coding.
Additional NCA control documents covering areas such as critical systems, cloud computing, data, and telework. They sit on top of the ECC and add requirements — they do not replace the baseline, a distinction often lost during scoping.
Assessment and scoring
Each framework scores differently. Using one framework's vocabulary for another is a common and costly confusion.
The SAMA CSF scoring method — six levels from 0 to 5, where level 3 is the minimum expected of supervised institutions. It measures how established and repeatable a control is, not simply whether it exists.
The CST CRF's tiered structure. CL1 covers foundational controls, CL2 adds more advanced requirements, and CL3 focuses on continuous monitoring and improvement. The applicable level is assigned by CST — it is not self-selected.
The ECC scoring approach — each applicable control is classified as implemented, partially implemented, not implemented, or not applicable. Exclusions require documented justification, which assessors question more often than inclusions.
The exercise of comparing current state against a framework's requirements to establish what is missing. Its output is only as reliable as the evidence behind each claimed status.
An entity's own evaluation of its compliance position, submitted to the relevant authority. Self-declared status without supporting evidence is the most common finding in Saudi regulatory assessments.
The ISO 27001 document recording, for all 93 Annex A controls, whether each applies, why, its implementation status, and the justification for any exclusion. Auditors examine it most closely, and it is most often found out of date.
Core GRC concepts
The distinctions in this section are what make multi-framework compliance possible. Collapsing them is where duplicated work originates.
What a framework asks for, in its own words and numbering. A requirement belongs to one framework and cannot be reused elsewhere — which is precisely why it is best kept separate from your controls.
What your organisation actually does to satisfy requirements. One control can satisfy requirements across several frameworks at once, which is where most of the saving in a multi-framework programme comes from.
The specific implementation activity beneath a control, with a named owner, a status, and a target date. This is where accountability actually lives — a department is not an owner.
The artefact demonstrating that a control operates. Evidence has a shelf life: a penetration test valid in March proves little in December, and an uploaded file is not evidence until someone has reviewed it.
A design artefact — a policy or procedure — shows a control was designed. An operating artefact — a log extract or completed review record — shows it actually ran. Most programmes we assess are strong on the first and weak on the second.
The link between a framework requirement and the controls that satisfy it. Doing this once is the step that removes duplication, and the only part of a compliance programme whose value compounds.
Whether a control applies to your scope, and the recorded reasoning where it does not. An exclusion without documented justification is a finding — assessors ask why far more often than whether.
Assets and systems whose disruption would have significant national impact. Private-sector organisations that own, operate, or host CNI fall within ECC scope regardless of sector.
Risk and audit
Every framework here mandates both. They are one loop, though most organisations we work with run them as two.
The likelihood and impact of a risk before any control is applied. It establishes the size of the exposure being managed and justifies investment in the control.
What remains once the linked controls operate. This is the number leadership decides on — and it is only credible if the controls behind it have actually been tested. A residual score with no tested control behind it is an opinion.
A formal decision to accept residual risk rather than treat it further. Legitimate when it carries a named owner, a stated rationale, and a review date. Risk accepted by default, because nobody decided, is a finding.
An identified failure or weakness raised during an audit or assessment. A useful finding records what failed, why it failed, who owns the correction, and when it is due — not merely that something was wrong.
The underlying reason a control failed, as distinct from the symptom. Fixing only the symptom is why the same finding reappears in a different system six months later.
A corrective action closing only when a verification step confirms it was completed and effective — not when the owner reports it done. Self-reported closure is the most common reason findings recur cycle after cycle.
Independent testing of whether controls genuinely operate. Mandatory under ISO 27001 clause 9.2, and present within the governance domain of the NCA ECC, SAMA CSF domain 3.2, and the CST CRF.
Privacy and personal data
Privacy is the only compliance discipline that runs on clocks you do not control.
The organisation that determines the purpose and means of processing personal data. The controller carries the primary obligations under the PDPL, including breach notification and responding to rights requests.
An organisation processing personal data on a controller's behalf. A processor moving data outside the Kingdom is still the controller's transfer, which is why processor mapping matters more than it first appears.
An inventory of what personal data you hold, where, why, on what lawful basis, with whom it is shared, and for how long. Every other privacy obligation depends on it, and it is the artefact most often missing when an investigation opens.
The legal ground relied on for a specific processing activity. It is established per activity, not per department or organisation — and where consent is the basis, it must be captured in a demonstrable form.
A request from an individual to exercise their rights over their personal data. Under the PDPL the response deadline is 30 days, and the clock starts on receipt — including through channels you did not design for it.
The obligation to notify SDAIA of a personal data breach within 72 hours. The clock starts on becoming aware — which happens at operational level, not when news reaches the privacy team.
Moving personal data outside Saudi Arabia. Conditions and any required assessment must be addressed before the transfer, with the reasoning recorded at the time rather than reconstructed when asked.
How long personal data is kept before deletion, defined per processing activity. A period defined in a policy but never enforced in a system is not a retention control — and retained data enlarges every breach that follows.
Definitions are the easy part
See what running these frameworks together in one connected system actually looks like.
