Where does our data get processed and stored?

d.AP runs in a European cloud environment. Either in your own account as a managed service, or with us, as SaaS.

Is d.AP GDPR-compliant and auditable?

Every answer is documented back to its source — role-based and granular. Auditability is part of the architecture.

How is role-based access control handled?

You define the permission concept. Technically, we can assign permissions down to the most granular level. In practice, employees typically see exactly what they would be allowed to see in the source systems anyway.

Which source systems can d.AP be connected to?

Through our federators, we connect to every existing interface technology (REST, JDBC, SOAP, etc.) and thus to all the systems you already have in place, such as ERP, CRM, MES, IoT platforms and document stores, etc.

Do we have to migrate our existing systems?

No. Source systems remain unchanged — d.AP federates across them.

Which LLM does d.AP use and can we switch it?

For Aluna - our AI agent - d.AP uses a model of our choosing that is hosted in Europe. At the same time, you can access d.AP with any agent & model deployed on your side and query the knowledge modelled within it.

How long does it take to reach a productive go-live?

Go-live with five to six sources typically takes three months. We start with the use case that offers the highest leverage.

How quickly can an additional data source be added?

Each additional source is added in one to two weeks.

Do we retain control over our knowledge model?

Yes, this is one of the most essential aspects of our platform. The ontology - and with it your knowledge - is modelled and stored in OWL, a worldwide open-source standard. This knowledge can be exported in that open format at any time and reused further.

What happens to our ontologies if we switch to another platform?

Your business logic is modelled in OWL, the worldwide open-source standard format, and can therefore be fully exported at any time and used in other platforms. You and your organisation's knowledge are not tied to our platform.

How is d.AP licensed?

Licensing is based on the sources and use cases you use, not on token consumption. That keeps costs predictable while usage grows.

Who operates the solution — us or digetiers?

Both are possible. At present d.AP runs in cloud environments only. Either in your environment, still managed and operated by us, or fully as SaaS - we then take care of everything. Important: in either case you retain full sovereignty over your data.

d.APThe AI Context Platform

AI understands everything. Except how you work.

d.AP models how your business works: facts, relationships and rules, applied by every agent, every time.

Modelled once, valid everywhere, yours to export.

Request a demo How it works

Why AI gets stuck in the enterprise

It's not just data that's distributed across the enterprise, the knowledge is too.

Your enterprise knowledge already exists. In ERP records, contracts, specifications, tickets and process documentation. What's missing, is the connection between them: which concepts mean the same thing, how they relate, which rules apply. That connective tissue lives in people's heads and in undocumented convention. To an AI, it is invisible.

An AI that only answers questions can work around this - with human support. An AI that acts on your behalf cannot. Autonomous agents need to know what a contract obliges, which customer record is authoritative, and which rule overrides which. That is the bridge d.AP builds

d.AP · Knowledge Layer in Action

Whatever the question, the answer will always be correct.

Aluna is d.AP's AI agent. Answers that are reliable — with full derivation, sources and explanation. No simple chat, no hallucinations.

Aluna answering a question about a plant outage in d.AP — with a supplier table, impact analysis and a recommended action.

Next step

Experience Aluna.

Request a demo View the architecture

Use cases

What you can do with d.AP.

Examples of how companies optimise their operational processes with connected knowledge that AI can use.

01 — Service

Solve tickets sooner

Classify a ticket without opening six systems — first analysis at a single point of access.

Learn more →02 — Sales

Master the variants

Sales, vehicle structure and importer data in one model — bill-of-materials views are queried, not requested.

Learn more →03 — Supply Chain

Catch supply failures

See instantly which suppliers, parts and routes are affected — before downtime starts costing you.

Learn more →04 — Quality

Trace defects back

Engineering, production and after-sales in a single view. Find root causes of defects faster.

Learn more →05 — Controlling

Steer beyond finance data

Real-time transparency across budgets and forecasts — grounded, reliable.

Learn more →06 — Revenue

Find potential in the base

Connect CRM, contracts and support — revenue potential in seconds.

Learn more →07 — Architecture

See what depends on what

Systems, processes and costs in one view — smarter IT, less risk.

Learn more →08 — Operations

Plan production realistically

A live view of orders, utilisation, defects. Agents right on the layer.

Learn more →09 — After sales

Spot churn before renewal

Contracts, charging sessions and support cases in one model — visible where a renewal is at risk.

Learn more →10 — Purchasing

Calculate commodity exposure

A commodity price rises: exposure calculated through to series, option variant and supplier.

Learn more →

Your use case

Those were examples. Perhaps you are facing an entirely different challenge?

All use cases

Architecture

Your systems stay. The knowledge becomes usable.

d.AP does not replace an operational system or a data lake. It lays a semantic layer over what you already have and makes enterprise knowledge accessible to every AI through a single endpoint.

What d.AP delivers

Secure connection to every internal and external data source.
Reliable usage Answers are based on your knowledge, not on statistical patterns.
Full provenance Every answer is traceable to source, logic and filters.
Immediately operational Enterprise-ready, incl. roles & permissions, SSO and scalability.

01
Consumers — Every application, one source.
Chats, agents, apps and BI all get the same view: MCP, A2A, REST.
02
Aluna — The knowledge agent.
The gateway to knowledge. Routes requests, authorises sources, delivers traceable answers.
03
Knowledge Graph — Your business logic, usable.
Your business logic as an ontology (RDF/OWL), modelled once, endlessly usable, always owned by you.
04
Federators — Connecting, preparing, mapping.
One federator per source: technical connection to any technology, data cleansing and mapping onto the ontology.
05
Source Data — Your operational reality — untouched.
Data products as well as operational systems (ERP, CRM, MES, etc.) and files — the living state of the enterprise.

Next step

Let's have a conversation about your situation.

Open product PDF

A semantic classification

Three things are called “semantic”. They don’t all mean the same thing.

BI Semantic Layer
ToolsdbtCube
StandardsOSISQL/PGQ
Query
Do all tools calculate the same metric identically?
A catalogue of queries: each entry defines which result a metric returns — “this query returns revenue”. Standardised as YAML on top of existing tables.
What gets defined is not the concept but a query against the raw data in its current shape. What “revenue” means has to be reverse-engineered from the SQL — the same backwards work as reading code whose intent was never written down. OSI standardises how those definitions travel between tools; it does not turn them into meaning.
Labeled Property Graph
ToolsNeo4j
StandardsGQLCypherSQL/PGQ
Connect
How is everything connected, so that I can traverse paths?
Data as nodes and edges with properties. Optimised for traversing relationships at speed.
The core point: it records how your facts connect — not how your concepts connect. Questions therefore stay patterns inside a bounded region of the graph, and a GNN predicts new edges from existing edges, from the facts. Meaning has to be recovered first — from labels, edges and properties, the same backwards work as in the semantic layer. It is expensive, and what ends up on top is a bolted-on ontology.
Knowledge Graph
StandardsRDFOWL
Understand
What do things mean — and which rules hold when machines act on them?
A formal model of your business terms (the ontology), populated with your real entities (the graph). Meaning, rules and data in one place.
Here meaning is not a reconstruction but part of the graph: classes, properties and rules sit next to the instances and are machine-interpretable. An AI does not have to guess what was meant — it reads it.

BI Semantic Layer

ToolsdbtCube

StandardsOSISQL/PGQ

Query

Do all tools calculate the same metric identically?

A catalogue of queries: each entry defines which result a metric returns — “this query returns revenue”. Standardised as YAML on top of existing tables.

What gets defined is not the concept but a query against the raw data in its current shape. What “revenue” means has to be reverse-engineered from the SQL — the same backwards work as reading code whose intent was never written down. OSI standardises how those definitions travel between tools; it does not turn them into meaning.

Labeled Property Graph

ToolsNeo4j

StandardsGQLCypherSQL/PGQ

Connect

How is everything connected, so that I can traverse paths?

Data as nodes and edges with properties. Optimised for traversing relationships at speed.

The core point: it records how your facts connect — not how your concepts connect. Questions therefore stay patterns inside a bounded region of the graph, and a GNN predicts new edges from existing edges, from the facts. Meaning has to be recovered first — from labels, edges and properties, the same backwards work as in the semantic layer. It is expensive, and what ends up on top is a bolted-on ontology.

Knowledge Graph

StandardsRDFOWL

Understand

What do things mean — and which rules hold when machines act on them?

A formal model of your business terms (the ontology), populated with your real entities (the graph). Meaning, rules and data in one place.

Here meaning is not a reconstruction but part of the graph: classes, properties and rules sit next to the instances and are machine-interpretable. An AI does not have to guess what was meant — it reads it.

We are proud to have supported these companies

Verband der Privaten Krankenversicherung

… and many more

Core elements of d.AP

It is always about: making knowledge usable.

d.AP turns enterprise knowledge into a form AI can interpret and execute directly.

01 — Modelling

Ontology Workbench

OWL Modelling · RML Mapping

A visual editor for modelling your world and linking it to the real data. So that this does not remain the domain of data engineers.

Learn more

02 — Referencing

Metrics Library

KPIs · Metrics · Evaluations

Arbitrary metrics, KPIs and evaluations can be generated automatically, stored in a library, and referenced from any external system through the API.

Learn more

Aluna chat: an answer to a question about a plant outage, with a table of the four affected suppliers and a recommended action.

03 — Analysing

Analysis Agent

GenAI · Aluna · Data analysis

Run complex data analysis in natural language as a business team — and be certain that the answers you receive are correct.

Learn more

An enterprise-wide knowledge layer built for the age of agentic AI.

Discover d.AP

Sovereignty

From Europe, for the whole world.

We build on European technology and open standards such as RDF out of conviction — without open technology there is no sovereignty.

Beyond that, our architecture is designed for the highest level of information security and European data protection.

Request a demo

Voices

The application delivers.

Voices
from deployments
of d.AP

A selection

01 / 02 — Manufacturing

Finally, an AI that does not hallucinate, and I see the source right there with it.

Enterprise ArchitectManufacturing

Use case Knowledge Layer for Service & Support.
Go-live with seven sources.

01Manufacturing

Finally, an AI that does not hallucinate, and I see the source right there with it.

Enterprise ArchitectManufacturing

Use case Knowledge Layer for Service & Support.
Go-live with seven sources.

02Automotive

The speed at which use cases were put into production is unique, and so is the ROI.

Head of Data & AIAutomotive

Use case Connecting five business domains into one integrated knowledge layer.

Further insights

Dive deeper.

Ontology5 min

Common questions

What decision-makers ask first.

Security, integration, time-to-value.
Choose a topic — or read through them all.

Your question is not covered? We reply personally.

Let's talk

You don't have to start from zero. Talk to us.

Every new agent, every copilot starts from zero — unless they share a knowledge layer. In 30 minutes we talk through your challenges and demonstrate d.AP live.

How it works

AI understands everything. Except how you work.

It's not just data that's distributed across the enterprise, the knowledge is too.

Whatever the question, the answer will always be correct.

What you can do with d.AP.

Solve tickets sooner

Master the variants

Catch supply failures

Trace defects back

Steer beyond finance data

Find potential in the base

See what depends on what

Plan production realistically

Spot churn before renewal

Calculate commodity exposure

Your systems stay. The knowledge becomes usable.