Why does AI need an ontology?

AI can now act in the real world, moving money, signing contracts, taking decisions, but it has no shared model of what is true. A database records that something happened; it does not capture what it meant, who was really behind it, or whether it was allowed. An ontology is that shared, machine-readable model of reality: entities, their relationships, and the rules that govern them. Without it, an autonomous agent cannot safely reason about the world it is acting on. KXCO provides that model.

What is the difference between a database, a blockchain and an ontology?

A database stores rows. A blockchain stores events on a shared, tamper-evident record. Neither understands relationships, and relationships are where meaning lives. An ontology is the layer above both: it defines what each entity is, how entities relate, and what proof each claim requires, so humans, institutions and AI can reason over the same world rather than store disconnected facts about it.

What is a knowledge graph, and how does KXCO use one?

A knowledge graph represents entities as nodes and their relationships as typed, directional edges. KXCO's ontology engine turns public sources, filings, official releases and reported coverage into a knowledge graph where every edge is a typed claim carrying a subject, predicate, object, source, as-of date and confidence, so exposure, ownership and dependency fall out of traversing the graph rather than reconciling spreadsheets.

What insight does an ontology give that a database or a dashboard cannot?

A dashboard answers the question you already knew to ask, because someone had to choose the columns first. An ontology holds the relationships, so the findings that exist only in the shape of the whole become visible. Pointed at the AI sector using public sources alone, the KXCO engine resolved six seed companies into 234 entities and 530 sourced relationships and surfaced three things no single filing contains: six of the world's largest companies converging on one firm that makes the lithography machine every leading-edge AI chip is printed on, fourteen circular flows where the investor is also the customer, and five supply chokepoints underneath an apparently diversified sector.

Why do investors, funds, public companies and regulators need an ontology?

Because every one of them is asking a version of the same question: what is this really, who stands behind it, what does it rest on, and what happens to me if that moves. An investor can hold two positions that look unrelated and depend on the same supplier, lender or jurisdiction. A fund needs look-through on demand rather than a quarterly reconstruction. A public company's biggest blind spot is its own ecosystem, the supplier behind the supplier and the dependency nobody wrote down. A regulator is looking for risk that concentrates between filings rather than inside any one of them. A lender needs the chain behind the collateral. All of that is graph work, done today by hand with documents that were never built to connect.

What does deploying an ontology across an organisation reveal?

The value compounds once it stops being a project inside one department. Pointed at customers, it produces one customer rather than one record per system: who introduced them, what they hold, who pays them, what they are exposed to, and where your own book is actually concentrated. Pointed at your people, roles, delegation, approval limits and access become part of the model, so who can commit the firm to what is answerable and key-person dependency is visible instead of anecdotal. Pointed at your ecosystem, it covers suppliers, partners, distributors and lenders and the dependencies they carry, including the concentration you inherit from other people's suppliers. Risk and opportunity are the same question asked in two directions, and both are properties of the graph.

Why should an enterprise ontology not live inside an AI model?

Because a model is rented and knowledge should not be. Train a model on your business and it has absorbed your knowledge rather than structured it: you cannot cite a weight, correct a single fact inside it, or take a customer back out of it, and a version change can move the answer with nothing to point at. If the only way to ask what you own, who you owe and what you are exposed to is a supplier's model, that supplier sits in the middle of your own understanding, on their pricing, retention and roadmap. And no supervisor, auditor or counterparty accepts a confident summary; they accept a claim with a source and a date that can be checked without trusting whoever is presenting it. KXCO delivers knowledge sovereignty and AI does not: the knowledge stays sourced, dated and portable, held by you, and the model is a component you can swap whenever something better arrives.

What is knowledge sovereignty?

Knowledge sovereignty is owning the model of your own business rather than renting access to it. The reasoning can be rented, since models improve and get replaced. The thing being reasoned over should not be, because it is the asset: identity, ownership, authority, history and the relationships between them, each claim carrying its source and the date it was true. Held as an ontology it stays auditable, correctable, portable and readable by any model. Absorbed into someone else's weights it becomes something you can only ask permission to consult.

Is the KXCO ontology a digital twin?

Yes, that is the most useful way to read it. Engineers do not run tests on a picture of a jet engine, they run them on a model that behaves like the engine. The ontology is that for a sector, an institution or a single instrument: the same entities, the same relationships, the same rules, and the same clock, since every claim carries a source and the date it was true. You can question it, stress a node and watch the effect propagate, which is what separates a working twin from a diagram.

Why do large language models give unreliable answers about a business without an ontology?

A language model is reasoning power with no anchor. Pointed at a business with no shared model of what is true there, it produces confident, well-written answers nobody can act on, because nothing in the system distinguishes a checked fact from a plausible sentence. Grounded in an ontology, every answer traces back to a claim, a source and a date, and a gap in the evidence shows up as a gap instead of being written over. It is the difference between software that can talk about a business and software that can operate inside one. Agents in particular cannot act on paragraphs: they act on entities, relationships and permissions they can read and write.

Can AI agents use the KXCO ontology?

Yes. AI agents are first-class users of the ontology, not bystanders. Humans, businesses, governments, machines and AI agents all reason over the same shared model of reality, with the same identity, signatures and permissions. The ontology is the intelligence layer; identity, signatures, permissions, post-quantum cryptography and blockchain are the layers that make every claim provable.

What is the trust infrastructure for the AI and quantum era?

It is the layer that lets autonomous software act in the real world without asking anyone to take a claim on faith. KXCO provides it: a shared, provable model of reality (the ontology) anchored by post-quantum cryptography (NIST FIPS 203/204/205, ML-DSA-65) on a tamper-evident record. KXCO is a software company and the software layer, never the operator; the licensed institution holds the licence and the custody.

What is the KXCO ontology?

It is a formal map of four primitives, Trust (quantum), Judgment (AI), Record (blockchain) and Permission (regulation), and the relationships between them. Nothing becomes real on KXCO until all four agree on it. Each KXCO product is an edge between primitives made into a working tool.

Is a token a thing or a claim?

A token is not a thing; it is a typed claim about a thing, living on a record anyone can check. Once the claim is public and machine-readable, asserting it is no longer enough, issuer, holder, backing and lifecycle must each be provable.

Why does tokenisation need an ontology?

Because a token is a typed claim, not a thing, and once that claim is public and machine-readable, asserting it is no longer enough. Issuer, holder, backing and lifecycle must each be provable, and an ontology is the shared definition of what each property means and what proof satisfies it.

What is proof-of-reserves look-through?

Look-through is decomposing any instrument into its full dependency graph, every claim, counterparty and piece of collateral, down to the reserve or real asset it ultimately rests on. It is computable, not spreadsheet archaeology, so exposure and coverage fall out of traversing the graph rather than a reconciliation.

How does post-quantum cryptography (ML-DSA-65) secure tokenised assets?

Trust is the mathematics that makes a claim unforgeable and keeps it unforgeable, even against a future quantum computer. KXCO realises it with NIST FIPS 203/204/205 post-quantum cryptography (ML-DSA-65 signatures), anchored on Armature L1, which is post-quantum from genesis. The proof is checked independently; the chain remembers.

What is KXCO's role, operator or software layer?

KXCO is the software layer, never the operator. The licensed institution holds the licence and the custody and issues the instrument; KXCO provides the machinery to prove the claim and the regulatory-intelligence layer to interpret the rules.

How does an ontology help with complex financial models?

Complexity in finance is composition, claims on claims on assets. An ontology keeps every relationship explicit, so an instrument can be mapped (look-through to the real asset underneath), reported on (exposure and coverage as a query rather than a reconciliation) and understood (reasoned over, with contradictions surfaced rather than hidden).

Does KXCO issue stablecoins or hold assets?

No. KXCO is the software layer, never the operator. The licensed institution holds the licence and the custody and issues the instrument; KXCO provides the machinery to prove the claim and the regulatory-intelligence layer to interpret the rules.

What is the four-agreement rule?

Nothing is real until trust, judgment, record and permission all hold at once. Remove any one and the token is forged, unreadable, deniable, or illegal.

Is KXCO the operator, or the software layer?

KXCO is the software layer, never the operator. The licensed institution holds the licence and the custody. KXCO holds neither.

Can provenance be backdated on KXCO?

No. Trust is forward-only: verification begins at go-live, and provenance is never backdated.

Is the proof stored on-chain or checked off-chain?

Off-chain proof, on-chain anchor. The chain remembers, and the proof is checked independently.

Is the regulator treated as an obstacle or a primitive?

The regulator is a primitive, not an obstacle. Permission is designed in, and kept live, cited and current.

Infrastructure for the Human–AI Economy

You HAVE the data, but can you SEE it?

The risk you never saw and the opportunity you never spotted are the same failure. The answer was already sitting in your business, and nothing showed it to you. What you own, who really stands behind it, what it rests on and what moves when that moves all live in the relationships between records, and neither a database nor a ledger keeps those. That is the missing layer, and it is the one your people, your regulators and now your machines all need. KXCO builds it: a working digital twin of the subject you need to understand, that you, your institution and your AI all reason over. The ontology.

"it's backed, trust us" → backing proven · holder identified · record kept · permission live

Seed names

234

Entities resolved

530

Relationships mapped

100%

Public data

Live · public data only
We pointed this engine at the AI sector using nothing but public sources. It resolved six names into 234 entities and 530 relationships, 465 of them carrying a source you can open, down to the five chokepoints the whole sector rests on and the one company at the root of them. That is what public sources alone allow. On your own assets, from the inside, the engine does not infer, it proves.

See it on public data →

Why you need this

Different seats. One missing layer.

Everyone in a market is asking a version of the same question. What is this really, who stands behind it, what does it rest on, and what happens to me if that moves. Nobody is short of data. Everyone is short of the shape of it, so today that question gets answered by hand, out of documents that were never built to connect.

Investors: what am I actually exposed toTwo holdings can look unrelated and depend on the same supplier, the same lender or the same jurisdiction. The exposure nobody can see is the one that reprices everything at the same time.
Funds and asset managers: look-through on demandConcentration, coverage, mandate fit and jurisdiction mix fall out of the model instead of out of a quarterly reconstruction. The same question, asked at any hour, returns the same answer.
Public companies: your own ecosystem is the blind spotThe supplier behind your supplier. The customer who is also your competitor's largest customer. The single dependency that nobody wrote down because no one team could see both ends of it.
Regulators and supervisors: systemic risk is a relationship problemRisk does not concentrate inside one filing, it concentrates between them. A shared model shows where it actually gathers, with every claim carrying the source it came from and the date it was true.
Banks and lenders: who am I really lending toFollow the guarantees, the parent, the pledge and the pledge on the pledge. Collateral is only ever as good as the chain behind it, and that chain is a graph, not a document.
Everyone else in the market: auditors, exchanges, insurers, counterpartiesAnyone whose job is to establish that something is what it claims to be is doing graph work right now, with phone calls and PDFs. The work does not change. The evidence stops being manual.

What data misses

Not in the rows. In the relationships.

Every fact below was already public. None of it was visible in any single filing, report or dashboard, because none of those hold the shape of the whole. We pointed the engine at the AI sector using public sources only. These are the things it found that nobody had a column for.

Six household names, one root.

Follow the dependencies down and six of the largest companies in the world converge on a single firm in Veldhoven that makes the lithography machine every leading-edge AI chip is printed on. It sits behind a Taiwanese fab, behind an American chipmaker, behind the names everyone actually watches. No filing states this. The graph does.

Fourteen loops where the investor is also the customer.

Capital goes out, revenue comes back, and each leg looks perfectly ordinary on its own line. Only the graph closes the circle and shows you how much of the sector's growth is buying its own supply.

Five chokepoints wearing a diversified disguise.

Read the rows and the sector looks broad and competitive. Traverse the relationships and the same handful of suppliers sits underneath all of it, which is a concentration no single balance sheet reports because no single balance sheet can see it.

A dashboard answers the question you already knew to ask. A model answers the one you didn't.

Open the engine and check all three →

Across the organisation

One model. The whole ecosystem.

The value compounds when this stops being a project inside one department. Point one model at your customers, your people and everything you depend on, and the questions that used to need a working group and three weeks become a traversal.

Your customers: one customer, not one record per systemWho introduced them, what they hold, who they pay, who pays them, what they are exposed to. Where your own book is concentrated becomes a property of the model rather than a study someone commissions.
Your people: authority, not an org chartRoles, delegation, approval limits and access sit inside the model. Who can commit the firm to what, on whose behalf, and what they touched. Key-person dependency becomes visible instead of anecdotal.
Your ecosystem: including the parts you do not ownSuppliers, partners, distributors, lenders and their dependencies. The concentration you inherit from other people's suppliers is real exposure, and it is missing from every dashboard you currently own.

Risk and opportunity are the same question, asked in two directions.

The digital twin

A twin, not a diagram.

Engineers do not run tests on a picture of a jet engine. They run them on a model that behaves like the engine. The ontology is that, for a sector, an institution, a counterparty or a single instrument. Not a diagram of the subject. A working copy you can question, stress and watch respond.

The same parts: entities, not tablesA person, a company, an agent, an asset, an obligation and an authority each exist once, defined once, and are referenced everywhere rather than rebuilt inside every system that needs them.
The same wiring: typed relationshipsOwns, controls, supplies, guarantees, is permitted by. Every connection carries its own type and direction, so the structure is something you can traverse instead of something you have to remember.
The same rules: constraints that biteWhat has to hold for a claim to stand is part of the model, not a note in a policy document. Break one and the twin tells you, which is how a contradiction surfaces before it becomes a loss.

And the same clock. Every claim in the twin carries a source and the date it was true, so you can tell what holds now from what held last quarter and from what was never checked at all. A twin of a sector, built from public data. A twin of your own institution, built from the inside.

Complexity, made legible

Map the models no one else can.

The dangerous instrument was never the complex one, it was the opaque one. Complexity is only composition: claims on claims on assets. An ontology keeps every relationship explicit, so you can hold the whole structure and still look straight through it, to what it truly rests on.

A twin of a tokenised fund, decomposed · hover or tab to look through to the real asset underneath

Every claim traces to something real and provable. Hover a node to follow the look-through.

Map: look-through, all the way downDecompose any instrument into its full dependency graph, every claim, counterparty and piece of collateral, down to the reserve or real asset it rests on. Computable, not spreadsheet archaeology.
Report: a query, not a reconciliationExposure, coverage, concentration, regulatory treatment and jurisdiction mix all fall out of traversing the graph, on demand, instead of a month of chasing the same numbers between teams.
Understand: reason over it, don't just store itStress one node and watch the effect propagate. Does the waterfall hold; is every leg permitted; is every claim proven. Contradictions surface instead of hiding.

Why now

Six shifts, at once.

None of this was true a few years ago. Together they broke the way the world keeps track of itself.

AI stopped advising and started acting.

Agents now move money, sign documents and take decisions directly, a new kind of actor the world has no record type for.

Value became programmable.

Money, assets and contracts are software now, a stablecoin, a central-bank currency, a tokenised building, each one a claim that has to be checkable.

Identity fragmented.

A person, a company or an agent is a different, disconnected record in a hundred systems that never agree on who is who.

Regulation went real-time.

What is permitted, by whom, and where is no longer paperwork after the fact, it is a live question asked on every transaction.

Trust stopped scaling.

You can no longer meet, or vet, everyone you transact with, let alone every machine acting on their behalf.

And our records got better at facts than at meaning.

Systems became excellent at recording that something happened, and no better at understanding what it was, or how it connects to everything else.

A database stores rows. A blockchain stores events. Neither understands relationships, and relationships are where meaning lives.

Why the machines need one

All that intelligence. No ground truth.

A language model is reasoning power with no anchor. Point it at a business and it will produce confident, well-written answers nobody can act on, because nothing in the system separates a checked fact from a plausible sentence. The ontology is what turns that reasoning into work. It is the difference between software that can talk about a business and software that can operate inside one.

Grounded, not guessed: every input carries a sourceAn answer built on the ontology traces back to the claim, the source and the date it was true. Where the evidence runs out, the gap shows up as a gap instead of being written over with something that reads well.
An agent needs more than a brain: reasoning is not reachNo agent can act on paragraphs. It acts on entities, relationships and permissions it can read and write. Without that layer it has judgment and no hands, and every action it takes is a guess about a world it cannot actually see.
The alternative is not none: a hundred private modelsEvery team already carries its own idea of what a customer, an exposure or an obligation is. None of them are written down and no two answer the same question the same way. One shared model is not extra structure. It replaces the hundred you are already running on.

Knowledge sovereignty

Rent the intelligence. Own the knowledge.

Every model you use is rented. Someone else's weights, on someone else's hardware, on someone else's release schedule. That is a perfectly good arrangement for the reasoning. It is a terrible one for the thing being reasoned over. Push your understanding of your own business into a model and you have handed over the asset and kept the invoice. KXCO delivers knowledge sovereignty. AI does not.

Weights are not a record: absorbed, not organisedTrain a model on your business and it has swallowed your knowledge rather than structured it. You cannot cite a weight, correct one fact inside it, or take a customer back out of it. Ask the same question after a version change and the answer can move, with nothing to point at and no way to tell whether you or the model was wrong.
Whoever holds the interface holds the asset: a vendor between you and your own businessIf the only way to ask what you own, who you owe and what you are exposed to is a supplier's model, that supplier sits in the middle of your own understanding. Their pricing, their retention policy, their availability, their roadmap. Keep the knowledge as data you hold and the model drops back to being a component you can replace.
Evidence, not recall: "the model said so" is not a filingNo supervisor, auditor, board or counterparty accepts a confident summary. They accept a claim with a source and a date attached, and the ability to check it without trusting the party presenting it. That is what an ontology holds. It is not something weights can produce, however fluent they sound.

Models will come and go. Your understanding of your own business has to outlive them.

So the model stays outside the knowledge, and the knowledge stays sourced, dated, portable and readable by whichever model you choose to point at it. Swap the reasoning whenever something better arrives. You should never have to migrate the truth.

The model

Four primitives. One system.

At the core of the ontology sit four roles, not four buzzwords: quantum, AI, blockchain and the regulator. Nothing becomes real until all four agree on it at once. That rule is shown by default below; hover or tab any node to trace how the pieces depend on one another.

Each product is an edge between two primitives · hover or tab a node to trace it

Primitive Live / active relationship Where a token becomes real

The stack

Everything else is an application.

KXCO is not a blockchain, a wallet or a signing tool. It is the layer those things run on. The ontology is the intelligence layer, the shared model everything above reasons over, and everything below makes provable.

ApplicationsWhat people and institutions actually useSurface

AI AgentsAutonomous users of everything below, a first-class actor, not a bystanderActors

The OntologyThe shared model of reality, meaning, relationships, reasoningIntelligence

IdentityWho and what everything is, person, company, agent, assetWho

SignaturesProof a claim was actually made, by the party who made itProof

PermissionsWhat is allowed, by whom, and where, checked liveAllowed

Post-Quantum CryptographyTrust that survives the machines that will break today's signaturesUnforgeable

BlockchainThe shared, tamper-evident record anyone can checkRecord

InfrastructureThe network and machines it all runs onGround

Humans, businesses, governments, machines and AI agents don't get five different views of the world. They reason over the same one.

The issuance domain

Every token is a typed claim.

And a claim everyone can now check. Each kind carries the same four edges, issuer permissioned, holder identified, backing proven, lifecycle recorded, that used to be a promise in a PDF.

Stablecoin: a claim on a reserve
reserve, proven, not asserted
redemption right, on record
issuer, permissioned
attestation, live
CBDC: a sovereign liability
central-bank issuer
legal-tender status
jurisdiction & policy controls
provable for decades, not years
RWA token: a claim on a real asset
underlying asset & legal wrapper
named custodian
provenance chain
valuation source

The axioms

What is always true.

The four-agreement rule.: Nothing is real until trust, judgment, record and permission all hold at once.
KXCO is the software layer, never the operator.: The licensed institution holds the licence and the custody. KXCO holds neither.
Trust is forward-only.: Verification begins at go-live. Provenance is never backdated.
Off-chain proof, on-chain anchor.: The chain remembers. The proof is checked independently.
The regulator is a primitive, not an obstacle.: Permission is designed in, kept live, cited and current.

Under scrutiny

How it holds up under scrutiny.

The questions that decide whether a model of reality is real infrastructure or just a diagram, answered plainly.

How the ontology is constructed: It is a shared schema of the entities that matter economically, identity, authority, assets, events, proof and history, and the relationships between them. Each entity is defined once and referenced everywhere rather than rebuilt inside every system. Where a claim has to be permanent, its proof is signed and anchored to Armature L1, KXCO's post-quantum record, so anyone can check it later without asking us.
How information becomes trusted: Nothing is trusted because KXCO says so. Every claim is cryptographically signed and recorded, so any party can check it independently. The signatures are post-quantum, NIST FIPS 203 / 204 / 205, so those proofs still hold decades from now.
Third-party extension: The ontology is a base others build on, not a closed product. Institutions and developers define their own entity types and relationships against the shared model, so their systems interoperate by default instead of through one-off integrations.
AI agent interaction: Agents act under a verifiable identity and delegated authority, reading structured facts and writing signed actions. Every agent action carries proof of who authorised it and exactly what changed, so autonomy stays accountable.
Network effects: Every new participant, asset and record makes the shared model more complete and more useful to everyone already on it. Because entities reference one another, value compounds as the graph grows, the more of reality that is modelled, the more that can be verified and acted upon.
Switching costs: Identity, history and proofs accumulate in the ontology and stay independently verifiable and portable, the record is not locked in a vendor silo. The real cost of leaving is re-establishing verified relationships elsewhere, which grows with the depth of history a party has built.

Go deeper in the engineering blog: how AI agents act on the ontology, and why BlackRock's quantum warning makes a shared, post-quantum model urgent.

KXCO does not ask you to trust the issuer's claim, or the issuer's reading of the law. It provides the machinery to prove the first, and one shared model, sourced and dated, to reason about the second.

See it on public data → See the products → Talk to us →