The operating model

One pattern that runs at every scale.

LuAir doesn't deliver reports about how your data should be managed. We install a working unit that manages it — seven capabilities, running on the data you care about most, proven on real work until your team operates it without us.

The unit of capability

A network of nodes, one orchestrator.

Every functioning data operation — a team, a department, an entire governance office — is the same unit at a different size. We call it a node. Nodes connect to an orchestrator that routes work, escalations, and status between them — and the orchestrator is itself a node, with the same seven capabilities inside.

The node network: an orchestrator in the middle, department nodes attached around it, the same seven capabilities inside every node Orchestrator a node itself Finance Operations Technology Marketing HR Legal & compliance INSIDE EACH NODE 01People & roles 02Intake 03Prioritization 04Escalation path 05Awareness 06Definition of done 07Communication cadence each node contains nodes — same pattern, smaller scope Every interaction flows through the orchestrator — and the orchestrator is itself a node with the same seven capabilities.
The pattern is recursive: the dots inside each node are its sub-nodes — teams within a department, individuals within a team — each running the same seven capabilities at its own scope. Understand one node and you understand the whole network, whether it's run by a person, a team, or software.
Two products

Build the node. Then run the services on it.

Everything LuAir does is one of two things: installing the capability, or delivering the work that runs on it.

Product 1, Enablement, builds the node; Product 2, Operations, runs the services on it PRODUCT 1 · ENABLEMENT Build the node Assess Plan Build Measure Graduate 1–3 people · ~8 weeks per node Your team graduates to run it without us Same backbone everywhere — custom internals PRODUCT 2 · OPERATIONS Run the services on it Issue resolution — one 7-step process Risk assessment Standards deployment Certification A defined catalog — ongoing after the node graduates Product 1 builds the node. Product 2 runs the services on it. Every team becomes a node that runs itself.
Product 1 installs the seven capabilities around the data that matters most, through a graduated engagement — we lead, then you lead, then we leave. Product 2 is the catalog of operational services any node can consume: yours, ours, or co-managed. Enablement always produces genuine walk-away capability; ongoing services are a choice, never a trap.
The universal process

Seven steps. Eight kinds of problem. One workflow.

Every data issue — a quality defect, an ownership dispute, a regulatory change — moves through the same seven steps. Only the questions change. One process to learn, one to measure, one to automate.

The seven-step resolution process 1 Identify 2 Understand 3 Document 4 Socialize 5 Agree 6 Review approval gate 7 Implement Steps 1–5 are led by the node. Step 6 is the review gate — nothing is implemented before approval. Step 7 executes, validates, and formalizes.
The eight problem types — coherence, quality, ownership, lineage, onboarding, compliance, regulatory change, retirement — all ride this same workflow, and aging work is never anonymous: every waiting item names the person it waits on.
Node today, agent tomorrow

Document a node and you've documented an agent.

An AI agent is instructions, skills, and tools. A node is roles, capabilities, and process. They are the same design at different levels of automation — so the playbook that installs your human operation is, word for word, the specification a governed agent can be built from. Nothing is bolted on afterward: the agent is born from the governance.

Each part of a documented node maps one-to-one to a part of an AI agent A NODE What you document Roles & decision rights Capabilities & catalogs The 7-step process Escalation paths AN AGENT What that becomes Instructions Skills & tools Control flow Human handoffs Same specification. The operator changes; the pattern doesn't.
Staff the node with people today; instantiate parts of it as agents when you're ready. Either way the accountability holds — every actor, human or AI, operates under a named owner, inside explicit decision rights, on an audit trail.
See where you stand

Seven questions map your operation to this model.

The readiness assessment scores nothing — it shows you the shape: which capabilities are absent, emerging, or already functioning.

Take the assessment Talk to us