The layer that turns many risk and protection domains into one system.
MAKOR connects policy administration, claims, partners and documents into a semantic picture. All core systems remain untouched, the AI in use receives reliable context — and every answer delivers verifiable evidence from the data.
Grown system landscapes keep valuable data out of the overall picture.
Your core systems have carried the business reliably for decades. Intelligent systems, however, can only read this data if they bring the right context with them.
Mainframe to rating
Each line of business carries its own history: life, health and P&C work on separate portfolios. Acquired and run-off portfolios additionally sit on legacy systems of their own.
File to payment
The facts of a case are in documents, the structure is in tables. Today the semantic bridge between the two exists in the head of the claims handler.
CRM to commission
A customer appears several times in several systems. The complete customer view is created manually today.
Classic integration projects tie up months of time and millions of budget before the first AI model goes into production. At the same time, the pressure to provide evidence under DORA, VAIT, Solvency II and the EU AI Act is rising.
How verifiably does an insurance company answer a question about its portfolio today?
Five stages describe how an insurer gets from manual information retrieval to a substantiated, reusable data foundation. In practice, individual lines of business — such as life, health or P&C — are often at different stages and therefore do not speak the same language.
Data states are formally checked for completeness and plausibility before they flow into decisions. The business meaning of the fields remains in the hands of the business units.
Completeness is established before release
Many firms start at A2–A3. The path from A2 to A4 determines the ability to provide information.
MAKOR starts at every stage.
Have your position assessedMAKOR docks on during live operation. Data sovereignty stays with the insurer.
No migration, no intervention in the core systems. MAKOR reads, models, links.
Ontologies as a universal language. Meaning and relationships between your data become visible.
Four principles that are not negotiable.
Meaning is more than data in tables.
The ontology understands what your data means and how it fits together: a claim belongs to a contract, the contract carries coverage clauses, the clause determines the obligation to pay, the expert report substantiates the facts.
Traceable AI answers. Every answer carries its context: you see which contract clause and which data point lead to the statement — explainable, verifiable, ready for decision.
Claim file logically linked to coverage clause and expert report — the obligation to pay becomes provable, reserves become more precise.
What matters is what runs live on real portfolios today.
| Acquired / monolithic platforms | MAKOR live | |
|---|---|---|
| Start | 6–18 months of lead time before the first model runs | 4 weeks to a functioning ontological picture |
| Understanding of data | Manual mapping per dataset, again when schemas change | The ontology models meaning once, new sources fit in |
| AI lock-in | Model dictated by the platform provider | Free choice of LLM, switchable at any time |
| Verifiability | Evaluations as a black box, no audit trail | Audit trail over every inference, ready to present to audit and supervisors |
| Ramp-up | Result on demo data in controlled environments | Result on your live portfolios, verifiable |
| Operations | Operation in external or third-country infrastructure | Developed and operated in Germany |
The proof on your data.
Connect existing systems
You choose the systems: policy administration, claims, mainframe or document archive.
AI of your choice
A freely chosen AI runs on the ontological picture of your live data.
Days to the result
A functioning ontological picture — live and verifiable.
What supervisors and auditors will want to see in future.
High-risk applications
Risk assessment and pricing in life and health insurance count as high-risk applications. Traceability of the decision and documentation of the data basis are required.
Demonstrate ICT dependencies
Evidence of data flows and dependencies in the ICT landscape is kept on an ongoing basis. A semantic picture makes these dependencies explicit.
Data quality and accountability
Supervisors expect verifiable data quality and documented responsibilities for each data stock.
Provenance and consistency
Actuarial models rely on portfolio data from several systems. The provenance and consistency of every figure become provable.
What insurers want to know before the first conversation.
What does MAKOR change in our core systems?+
MAKOR reads from the source systems and places a semantic layer above them. Your policy administration and your mainframe remain in their existing state.
Where does our data run?+
MAKOR is developed and operated in Germany. Operation in your own data centre is possible.
Which AI are we tied to?+
You choose the model and can switch it freely later.
What does the audit trail look like in concrete terms?+
Every AI conclusion (inference) — that is, every answer the AI derives from your data — is logged and archived together with the data points used. The evidence is ready to present.
What do we need to provide for the 4-week PoC?+
Read access to two or three systems, a business contact person and a delimited section of a process.
How do we start if we are at stage A1?+
An initial assessment shows your biggest gaps. The PoC starts with a section that quickly brings clarity.
15 minutes are enough for a first impression.
We show you live how a semantic picture of portfolio, claims and documents is created — and what your AI can answer from it.