Enterprise technology operating standard · Draft v0.1

Enterprise technology,
re-industrialised.

Meridian replaces governance theatre with an AI-native control system: persistent crews, paved roads, bounded agents and continuous proof. Move at the speed your risk allows—without choosing between delivery and assurance.

01Flow

Pull small valuable passages through paved roads.

02Proof

Generate regulatory and engineering evidence by construction.

03Agency

Give AI authority only where behaviour is bounded and observable.

04Joy

Design out status recital, form filling and avoidable cognitive load.

The historical lens

Industrial revolutions scale when the system around the machine changes.

Infrastructure

Common utilities multiply local invention. Meridian makes platforms the enterprise utility layer.

Standards

Shared interfaces let independent actors coordinate. Meridian standardises boundaries and evidence, not every local choice.

Measurement

Instrumentation turns complex systems into controllable systems. Meridian makes telemetry part of governance.

Mechanisation

Machines absorb repetition; people move toward judgement and new craft. AI follows the same pattern.

Modularity

Small replaceable parts accelerate improvement and contain failure. Enterprise architecture should do the same.

Organisational redesign

Electric motors did more when factories stopped imitating line shafts. AI does more when work stops imitating ticket queues.

“Flow with proof.” The Meridian design constraint
The operating system

Four planes. One digital thread.

Separate intent, reusable platform capability, delivery work and assurance—but connect every object so evidence moves with the work.

01

Intent Plane

Outcomes, strategy, risk appetite, obligations, capital and Missions.

02

Platform Plane

Self-service identity, cloud, delivery, observability, data, AI and policy.

03

Work Plane

Persistent Crews discover, build, run and improve enduring services.

04

Proof Plane

Controls, evidence, evaluations, exceptions, assurance and replay.

THE DIGITAL THREAD
ObligationControlServiceChangeEvidenceOutcome

Ask questions across the whole system instead of reconstructing answers from spreadsheets.

The human system

Crews own outcomes.
Stewards encode judgement.

Persistent multidisciplinary Crews own services end-to-end. Security, privacy, architecture, resilience and risk become Stewardship Circles whose goal is to turn recurring review into policy, patterns and paved roads.

CrewBuild + run an enduring service or platform
StewardNamed accountable human for a material object
CircleEncode cross-cutting expertise without becoming a queue
GuildGrow craft and social learning across Crews
SwarmTemporary coalition for a cross-cutting Mission
AI as enterprise utility

Not a copilot. A bounded second workforce.

Agents can scout, map, make, verify, observe and operate. Authority is granted to a task and environment—not inherited from a model brand.

Scout

Research, precedent and change sensing.

Cartographer

System, dependency, data and regulation maps.

Maker

Code, IaC, tests, migrations and runbooks.

Verifier

Evals, policy checks and evidence challenge.

Sentinel

Reliability, security, cost and model behaviour.

Operator

Bounded actions under runtime policy.

Normative core

A standard you can query.

Search the current Meridian clauses by domain, ID or concept. The goal is stable semantics that can be turned into policy and evidence—not another 400-page shelf document.

Practice library

Meetings should be worth attending.

No mandatory stand-up, sprint ritual or programme increment. Use practices when they create decisions, learning, confidence or human connection.

Applicability compiler

One operating system. Many regulatory profiles.

Toggle characteristics of an organisation to see which reference profiles become relevant. This is a design aid—not legal advice or a substitute for formal applicability analysis.

ACTIVE REFERENCE PROFILES
The 100-year bet

Tools will change. The invariants should not.

Purpose. Small reversible change. Explicit interfaces. Observable systems. Human accountability. Tested resilience. Reusable proof. Continuous learning. These survived steam, electricity and computing—and they are strong bets for whatever follows AI.

Read the complete Meridian Standard