Services
We build trustworthy data systems.
Systems that turn complex, fragmented data into decisions you can rely on – and defend.
We build everything from data pipelines for financial and climate-risk analysis to systems that verify project status, evidence and reporting against source documents.
That means reliable data collection, clear structure, rigorous curation and transparent lineage from the original source to the final metric. Because when decisions matter, a number is only as valuable as your ability to prove where it came from.
Risk and reporting for finance
Some of the hardest data problems show up in finance, where a number does not just need to be right – you need to be able to explain and defend it.
An ICAAP submission. A Pillar 3 disclosure. A climate-risk figure a supervisor is going to ask about.
Often, the number comes from a vendor model, a spreadsheet or a chain of calculations that nobody fully owns anymore. The result is there, but the path to it is difficult to reconstruct.
We rebuild that path as a system you can inspect.
From raw hazard and exposure data through to monetary risk metrics such as EAD, VaR and CVaR, we preserve the intermediate steps, source indicators, assumptions and limitations alongside the result. When someone asks how a number was reached, you can show them step by step.
We build these systems for banks working on ICAAP and regulatory reporting, insurers analysing portfolio risk, and asset managers assessing assets and portfolios.
Other systems we build
We apply the same approach to other evidence-heavy problems: build the system so every important output can be traced back to its source.
Examples include:
- Verifying reported project or programme status against independent evidence.
- Reporting and compliance data pipelines with full lineage back to source.
- Evidence-based AI and knowledge systems that ground every output in a source and check it against the rules.
Data pipelines you can audit
When reporting data is scattered across issuer PDFs, portfolio files, filings and internal systems, producing a number is only half the job. You also need to prove where it came from.
We build data pipelines that preserve that link – down to the source document, page and passage where needed.
What comes out is verifiable data, not another opaque score.
Independent verification
Sometimes a reported claim needs to be checked against evidence that did not come from the organisation making it.
A project reported as complete. An allocation reported as spent. An asset reported as compliant.
We build systems that compare those claims with independent evidence, including geospatial and satellite data where appropriate, and show where the evidence supports the claim – and where it does not.
AI for regulated processes
AI can be useful in regulated work, but only if its outputs can be inspected and challenged.
We build AI systems that connect answers to their underlying sources and apply explicit rules and validation checks before those answers enter a business process.
Depending on the problem, that can combine retrieval, structured knowledge and deterministic validation.
Where the methods come from
Some of the methods behind our client work are developed and tested through applied research projects under Austrian and European programmes, including FFG, Horizon Europe and ESA.
That work gives us room to solve difficult methodological problems, test new approaches and validate them before we use them in production environments.
Research is not separate from the commercial work. It is one of the places where we develop methods that later become reliable systems for clients.
How we handle your data
Your data stays yours.
We use only what the project requires. Where necessary, we can work entirely within your infrastructure so that sensitive data never leaves your environment.
We work under NDA and handle personal data in accordance with GDPR. Access is limited to the people directly involved in the project.
At the end of the engagement, we return or delete project data according to your requirements.
We do not share your data, and we do not reuse it for our own purposes.
Last updated
Tell us what you’re working on
An asset. A portfolio. A dataset. A reporting process. A system that needs to be built – or needs to stand up to scrutiny.
Tell us what you’re working on, and we’ll tell you whether we’re the right people for it and how we would approach it.