Your industry's language, encoded for AI
We start from the open standards your sector already uses, extend them with your business meaning, and add the sensitive-data rules your regulators expect.
Ship agents enterprises will actually approve
If you build AI products for regulated customers, their security review will ask what your agent can see, why, and how you prove it. We give your product an ontology-backed context and policy layer — a reusable semantic contract between your agents and each customer's data.
Core classes
Tenant · Agent · Tool · ToolScope · Dataset · DataClass · Purpose · Delegation · Decision
Discuss your ai use caseCompetency questions the ontology must answer
- Which tools can this agent call on behalf of this user, in this tenant?
- Which datasets contain data classified as personal or confidential?
- Which purpose was declared when this fact entered the context window?
- Which decisions relied on data from a now-revoked connector?
Example data guardrails
Claims, underwriting and service agents with boundaries built in
Insurance data mixes policy terms, party identities, health details from bodily-injury claims and financial information. We model parties, policies, coverages, claims and loss events so that every agent — adjuster copilot, vendor portal, fraud triage — sees exactly what its role and purpose permit.
Core classes
Party · Policy · Coverage · Claim · LossEvent · Vehicle · Property · Reserve · Vendor · Adjuster
Discuss your insurance use caseCompetency questions the ontology must answer
- Which coverages on this policy apply to this loss event?
- Which open claims involve a party linked to a prior SIU referral?
- What evidence supports this reserve recommendation?
- Which facts may a repair vendor's agent receive for this claim?
Example data guardrails
Clinical and operational AI that respects minimum necessary
Healthcare AI must reason across patients, encounters, conditions, medications and providers while honouring consent and care relationships. We align to HL7 FHIR resources and clinical terminologies so care-team copilots, prior-authorisation agents and research assistants each get the minimum necessary context.
Core classes
Patient · Encounter · Condition · Observation · MedicationRequest · Practitioner · CareTeam · Consent
Discuss your healthcare use caseCompetency questions the ontology must answer
- Is this practitioner on the patient's active care team?
- Which observations support this prior-authorisation request?
- Which records are excluded by the patient's consent directive?
- Which cohort records are de-identified enough for this research purpose?
Example data guardrails
Research, risk and client-service AI grounded in FIBO
Financial institutions need agents that understand instruments, issuers, legal entities, holdings and events — and that respect information barriers, client confidentiality and suitability rules. We ground the model in FIBO and your internal golden sources.
Core classes
LegalEntity · Issuer · Instrument · Holding · Portfolio · Account · Client · CorporateEvent · InfoBarrier
Discuss your financial services use caseCompetency questions the ontology must answer
- Which portfolios hold instruments issued by entities affected by this event?
- Which clients' data sits behind an information barrier from this desk?
- What is the ownership chain from this account to its ultimate beneficial owner?
- Which source filing supports this exposure figure?
Example data guardrails
Agents that understand assets, parts and process — and protect IP
Manufacturing knowledge is spread across PLM, ERP, MES, maintenance logs and supplier contracts. We connect equipment hierarchies, bills of materials, suppliers and process steps so maintenance, quality and supply-chain agents reason across them while proprietary recipes and supplier pricing stay protected.
Core classes
Site · Line · Equipment · Component · Part · BOM · Supplier · WorkOrder · ProcessStep · Recipe
Discuss your manufacturing use caseCompetency questions the ontology must answer
- Which finished products depend on this single-source supplier?
- Which equipment shares the failing component and when was it last serviced?
- Which process steps changed before this quality deviation?
- Which recipe parameters may a contractor's agent view?
Example data guardrails
Threat intelligence and cyber agents with need-to-know
Security operations demand fast correlation across assets, vulnerabilities, identities and threat intelligence — and strict control over who learns what. We model the cyber domain with STIX and ATT&CK concepts and enforce classification and need-to-know on every agent.
Core classes
Asset · Identity · Vulnerability · Indicator · ThreatActor · Technique · Incident · Control · Classification
Discuss your security use caseCompetency questions the ontology must answer
- Which internet-facing assets are exposed to techniques used by this actor?
- Which controls mitigate the techniques seen in this incident?
- Which indicators are TLP:RED and must not leave the response team?
- Which identities had access to the compromised asset in the last 30 days?
Example data guardrails
What every domain ontology includes
Domain core
Entities, relationships and rules that answer your competency questions.
Sensitivity & purpose
DPV-aligned data classes, purposes, consents and legal bases.
Provenance
PROV-O lineage from every fact to its source system or document.
Shapes & tests
SHACL constraints and executable competency-question tests.