The consumer

The model is not what the agent reads.
It is what the agent stands on.

The metamodel gives an agent a grammar it cannot step outside of. The instance graph gives it facts it did not write. With both, answering a question about your organisation stops being an act of composition and becomes an act of traversal — and the difference shows up in the one place that matters: whether you have to check the answer before you forward it.

Retrieval and traversal

Retrieval returns what resembles your question. A traversal returns the path to the answer.

Both start the same way: find the thing being asked about.

retrieval

stops there

Stops there, and hands the nearest paragraphs to the language model.

traversal

starts there

Starts there. The node is the entry point, and the relationships leaving it are the rest of the work.

Ask what a supplier failure reaches, in the Adventure Works model:

typed traversal · three hops · four types, three departments, one query

Supplier

Victory Bikes

supplies →

Product SKU

Touring Tire Tube

depends on ←

Sales Order

SO70282

is placed by →

Retail Shopper

Dalton Adams

Three hops, and the agent inferred none of them. It did not work out that a tyre tube belongs in a bicycle or that an order implies a customer — each hop is a typed relationship somebody or some system already asserted. The chain crosses a supplier, a product SKU, a sales order and a person: four types, three departments, one query.

And one hop further out is the part a paraphrase never reaches. The same component either has a second agreement or it does not — this one does not: one agreement, and no other — so the answer is a single point of failure, named. For the tyres the same supplier also provides, the traversal returns a second agreement instead, and the answer becomes an impact with a mitigation attached. That difference is not judgement. It is the second edge on the same node, present or absent.

Why the answer holds

The language model does not work out the answer. It reads the question and reports what the traversal returned.

Three jobs, and only the first and the last are language:

01 · language

Translate the question into a traversal

Which node to start from, which relationship types to follow, how far, and what counts as a hit. This is the part language models are good at.

02 · deterministic

Execute it deterministically

The traversal runs as a graph operation over typed edges. Same question, same model, same path — twice, and next quarter.

03 · language

Narrate the result

Put the finding in a sentence, with the nodes and relationships it came from still attached to it.

Organisational intelligence is not a language problem with a graph behind it. It is a graph problem with a language interface — and asking a model to reason about impact from prose is asking it to do the one thing it cannot be held to.

Three things follow, and none of them is a prompt.

what you can check

Every claim in the answer is an edge. Disagreeing with the answer means opening a node, which is a different kind of argument than disagreeing with a paragraph.

what gets recorded

The path is the audit. What the agent saw, and when, is a query against the same model — not a log in a second system.

what improves it

Accuracy tracks the model, not the prompt. Another typed relationship is another question answerable tomorrow; a better prompt is not.

At the edge of the model

Where the model is silent, the agent says so — and says which silence it is.

Asked who else could cover the purchase orders assigned to one buyer in Purchasing, the agent in the demo tenant did not answer that nobody could. It answered that there were

no secondary co-assignees modeled in the graph.

Read the qualifier. It is not a hedge — it is the whole claim. There may well be a colleague who covers her every August; what the model states is that nobody has recorded it. One of those sentences you can act on, and the other you would have had to verify before you did.

An agent over documents cannot make that distinction, because there is nothing for it to be silent against. Absence in a document store is indistinguishable from a page nobody wrote. Absence against a type is a finding: the relationship exists in the grammar, this node does not have it, and that gap has an owner — because the responsibility matrix belongs to the type, so every node knows who answers for what is missing on it.

Which is also the honest answer to the fair question about any model: what if ours is incomplete? It will be. The agent will tell you where.

Unprompted

Answering is the part you have to start. Watching is not.

A typed graph with state is something that can be monitored, and monitoring is where an agent earns a standing seat rather than a chat window. It notices:

that a process lost its owner

that an approver holds a mandate they have never used

that hundreds of orders now route through one person with no co-assignee recorded

that a skill has exactly one holder on the evening shift

None of those is a question anyone would have thought to ask on the day it became true. All of them are states the model can be checked against — and each one routes to a named person, because the responsibility matrix is a property of the type and not a notification rule somebody configured.

Autonomy of attention, not of authority. The agent identifies, reports, monitors and escalates. It does not change your organisation, and the boundary is not a policy in a prompt: what it is able to assert at all is limited by a grammar it did not author and cannot extend.

One more objection

One more objection, and it is the right one.

So far this describes a model that is true today. You have seen this part fail before — the catalog nobody updated, the CMDB that was wrong within a year. That objection deserves its own answer.

the answer

A model that stays true

Half the model is read from your systems and cannot drift. The rest has an owner, and the owner gets asked.

A model that stays true →

or check it yourself

Open the demo

A fully modelled organisation, one click with your work account. No form, no call.

Open the demo