


THE PRECEPTS
Our name is the Latin for the rules. These are ours.
We publish them because a standard you have not written down is a preference, not a standard. And because a client should be able to hold a practice to something more specific than a claim of integrity.
Nine precepts. Nine letters. If we breach one, tell us.
P — Primary sources only
We cite the regulator's own published text, in the binding language where one is specified, with a verification date on the page. We do not cite law-firm summaries, vendor blogs, consultancy commentary or aggregator sites as authority for what a regulation says. Those sources are often useful and sometimes excellent, but they are secondary, and a secondary source is not evidence of a regulatory obligation. This matters more than it sounds. The regional market now contains a great deal of confidently-written material about GCC and African frameworks that cannot be traced to any issuing authority, some of it machine-generated, much of it internally inconsistent on dates and version numbers. We have found instruments described as current that had been repealed, and compliance deadlines asserted with no gazette citation behind them. Where sources conflict and we cannot resolve the conflict against a regulator's own publication, we tell you that instead of choosing the convenient answer. You will see it on our own pages: statements that we cannot confirm a version or a status, with an instruction to verify with the authority and with counsel. That is not hedging. It is the only honest position available when the primary source does not say.
R — Requirements before tooling
A control you cannot trace to a requirement is a cost you cannot defend. We design first and procure second, always in that order, including when it is commercially inconvenient for us. If the right answer is a topology change rather than a purchase, that is what we will tell you — and it is frequently the most valuable conclusion an architecture engagement can reach. Expect a practical consequence. Early in an engagement we spend time on requirements decomposition that produces no visible security improvement. Clients sometimes find that uncomfortable. It is the phase that determines whether everything after it is defensible.
A — Architecture over assessment
A maturity score tells you how you compare. It does not tell you what to build, and you cannot remediate a number. The region has a large supply of gap assessments. What it has very little of is architecture — conceptual, logical and physical design, with trust zones, data domains, flow topology, key custody and enforcement boundaries, that an engineering team can build from and a regulator can be walked through. We will produce an assessment where you need one. We will not produce one without an architecture to go with it, because a finding without a design is a problem transferred back to you with a report attached.
E — Evidence as the deliverable
Compliance is not a state you reach. It is a claim you must be able to substantiate on any given Tuesday, about an estate that changed last week. A claim requires evidence. Evidence requires an unbroken thread from the clause that binds you, to the control that answers it, to the artefact that demonstrates the control is working. Building that thread is the work. In most organisations, evidence is assembled by a small number of people, manually, in the weeks before an assessment, from systems never designed to produce it. It usually works. It is expensive, fragile, dependent on individuals, and it tells you nothing about your position in the eleven months of the year when nobody is asking. We design evidence generation into the controls. The test we hold ourselves to: when an inspector asks how you satisfy a specific requirement, the answer should take a query, not a quarter.
C — Conflicts surfaced early
Organisations operating across multiple jurisdictions routinely hold requirements that cannot all be satisfied by a single architecture. Some of those conflicts are irreducible. No amount of design resolves them. What can be done — and what is almost never done — is to identify them in week two rather than month twenty, put them in front of the people accountable for the trade-off, and let that decision be made deliberately with the cost visible. A conflict found early is a strategic choice. The same conflict found during an inspection is an incident with someone's name on it. We will tell you about conflicts even when the conflict means the engagement you asked for is not the engagement you need.
E — Earnings declared
We hold declared technology partnerships. We publish them, we keep the list deliberately short, and we state what we earn on anything we recommend, in writing, before we recommend it. Assurance is firewalled from resale. On independent assurance work — architecture review, framework assessment, audit-adjacent work — we take no resale margin, referral fee or rebate on any product named in those findings. If you later choose to buy a recommended product through us, that is a separate decision you make freely, with our interest disclosed. It is never a condition of the assurance work. Findings are not for sale. No vendor has any right of review over our findings. No vendor target, quota or tier requirement influences a severity rating or an architectural recommendation. One alliance per category. We hold a single data security posture management partnership, deliberately, and intend to keep it that way. Any client or partner may ask at any time for written confirmation of our financial interest in anything we have recommended. We will provide it. Our full position is published at praecepta.co/declaration-of-interests.
P — Principal-led
The architect who scopes your engagement designs the architecture and presents to your board. There is no pyramid, and no handover to a graduate once the contract is signed. This is a constraint on us as much as a promise to you. It means we take on fewer engagements than we could sell. We think that is the correct trade. Where an engagement needs more hands than one, we bring named associates on pre-agreed terms, and we tell you who they are and what they will do before you sign — not after.
T — Traceable by design
Every control we deploy traces forward to the evidence it produces and backward to the clause that demanded it. Every significant architectural decision is recorded: the decision, the alternatives considered, the requirement it serves, the trade-off accepted, the cost of reversing it, and the date. That record is the difference between an architecture and a diagram.
A — Accountability retained
Accountability for your organisation's security sits with your executive and your board. We inform it, sharpen it, and make it defensible. We cannot and do not assume it, and any adviser who implies otherwise is selling something that will not be there when it is needed. We also write our own exit into the engagement. A good engagement has a defined end state — a completed architecture, a transferred capability, a full-time hire. An adviser with no exit in mind has a commercial interest in your dependency. Registers, decision records, dashboards and runbooks are handed over in formats your team can maintain.
Holding us to them
These precepts are a statement of our own standards. They become contractually binding when incorporated into a signed engagement letter.
If you believe we have breached one, tell us: ops@praecepta.ae. We will respond in writing.
Last reviewed: November 2026. Reviewed annually.