EBA ITS 2023/2083 · Ontology-driven validation

The compliant data backbone for the NPL secondary market.

Ingest, structure and validate non-performing loan packages against EBA ITS 2023/2083 — every field checked against a formal ontology and an auditable SHACL ruleset. Built for originators, servicers and credit buyers.

EBA 2023/2083 compliantSHACL ontology engineEnd-to-end NPL lifecycle
PORTFOLIO · NPL-2024-RT07
Compliance report
review_ready
83.7%
overall EBA ITS score
counterparty.legal_entity_idALWAYS98%
collateral.property_valueCONDITIONAL71%
enforcement.court_statusMISSING0%
ruleset · eba-its-2023/2083SHACL validated
FOR INSTITUTIONS

Originators, servicers and buyers: standardize messy NPL data, prove EBA compliance, and trade with confidence.

See the platform
FOR CREDIT BUYERS

Run due diligence on verifiable, EBA-validated portfolio data before you commit.

Learn more
THE MANDATE

Not a new idea — Europe already prescribed it.

In 2017 the ECB diagnosed the euro-area NPL secondary market as failing — opaque, fragmented, illiquid — and prescribed transaction platforms whose first, essential layer is harmonised data and independent validation. The EU Council tasked the EBA with the data standard, now the EBA ITS (EU) 2023/2083. NPLEX builds exactly that layer.

Harmonised data templates for loan tapes are an essential element of the platform. … Independent validation of the reported NPL data would be a key function.

ECB Financial Stability Review, November 2017
€373bn
EU/EEA NPLs still on bank balance sheets
EBA Risk Dashboard, Q3 2025
~€110bn
new NPLs flowing in every half-year
EBA, H1 2025
67 / 60%
investors active in the market — ten take 60% of volume
ECB FSR, 2017

The stock has fallen from €1.3tn (2017) to €373bn (2025) — but the failures the ECB diagnosed are structural: heterogeneous data, opacity, and a handful of buyers. With ~€110bn of new NPLs each half-year, a trusted data layer is a permanent need, not a cyclical one.

The cost of the missing layer is measured, not theoretical.

A bad-loan overhang doesn't just sit on a balance sheet — it throttles new credit to the real economy. The bottleneck is the data: heterogeneous and unverifiable, it keeps the secondary market fragmented and priced at a discount, so banks carry impaired books for longer.

Banks themselves report that NPL ratios tighten their credit standards — in the ECB Bank Lending Survey, now asked every quarter.

ECB Bank Lending Survey — ad-hoc question, quarterly since 2025
Sources: ECB Financial Stability Review (Nov 2017) · EU Council NPL Action Plan (Jul 2017) · EBA Risk Dashboard (Q3 2025) · EBA ITS (EU) 2023/2083
THE NPL LIFECYCLE

One end-to-end pipeline, from raw package to transaction-ready portfolio.

01
Ingest
Drop packages in any format and language.
02
Structure
AI maps raw data to EBA 2023/2083 fields.
03
Validate
Field-by-field compliance against the SHACL ontology.
04
Manage
Navigate portfolios, loans, counterparties, collateral.
05
Transact
Hand over auditable, EBA-validated portfolios to counterparties and regulators.
PLATFORM

The data and compliance infrastructure, in three products.

01 / INGESTION

AI Ingestion

Turn unstructured NPL packages into compliant portfolios. Upload .zip/.xlsx/.json/.txt, run AI extraction and EBA field mapping, then review and approve.

PIPELINE STATUSreview_ready
Extraction100%
EBA field mapping100%
02 / COMPLIANCE

Compliance Check

Score every portfolio against EBA ITS Annex I, template by template and position by position. Export auditable reports.

83.7%By Template
T1 · Counterparty96%
T3 · Collateral74%
T5 · Enforcement58%
03 / PORTFOLIO

Portfolio Navigator

Manage portfolios, loans, counterparties, collateral and enforcement with canonical EBA data and live compliance.

12
PORTFOLIOS
€2.4B
TOTAL GBV
86%
AVG COMPL.
RT07 · Retail
CM03 · SME
DATA SOVEREIGNTY

Your data never leaves you.

NPLEX runs on your own infrastructure. Your NPL data is validated in place — we never see it, store it, or transmit it. You hold the data; we provide the software.

01

Runs on-premise

Install NPLEX inside your own systems. Portfolio and borrower data stay entirely within your environment — consistent with GDPR and bank-secrecy obligations.

02

Formal validation gate

Before a snapshot can be finalised, the ontology and SHACL ruleset verify the data against EBA ITS 2023/2083 — every issue traced to its field code.

03

Portable, encrypted snapshot

A validated snapshot is sealed into an encrypted .nplex file that you exchange directly and privately with your counterparty. No intermediary sees the contents.

We provide the software. You provide the infrastructure. The data is yours.
COMPLIANCE & REGULATORY

Built on the EU regulatory corpus.

CANONICAL STANDARD — VALIDATED AGAINST

EBA NPL templates · Reg (EU) 2023/2083

The ontology and SHACL ruleset are built field by field from the EBA NPL data templates — Commission Implementing Regulation (EU) 2023/2083, implementing Article 16(1) of Directive (EU) 2021/2167. This is the single standard we validate against; where a cross-referenced framework diverges, the EBA NPL rule wins.

CROSS-REFERENCED FRAMEWORKS (MODELLING CONTEXT)

Directive (EU) 2021/2167

Credit Servicers & NPL Secondary Markets Directive — its Article 16(1) mandates the EBA data templates.

AnaCredit · ECB

ECB credit dataset (Reporting Manual, Part II) — counterparty and instrument attribute definitions.

ESMA RTS · 2020/1224

Securitisation disclosure RTS (NPEL templates) — e.g. the percentage-point value convention.

FINREP · 2021/451

Supervisory-reporting ITS — counterparty sector classification (Annex V).

IAS/IFRS · 2023/1803

EU-adopted accounting standards (IAS 1/7/16/32, IFRS 16) — the counterparty financials block.

CRR · 575/2013

Capital Requirements Regulation — prudential definitions (e.g. annual turnover, Art. 153(4)).

THE DATA FOUNDATION

A formal model of the EBA standard — not a spreadsheet.

NPLEX is built on an OWL ontology and a SHACL ruleset that encode the EBA NPL data templates field by field. Validation is declarative and deterministic: every check is traceable to its EBA field code, and every report is stamped with a versioned ruleset.

119
EBA data fields modelled
199
SHACL constraints across 112 shapes
10
entity types, 14 typed relationships
5
EU frameworks cross-referenced
5 EBA templates · 4 validation profiles (draft → save → snapshot → consistency) · every report carries its ruleset version hash.

See your NPL portfolios become compliant and transaction-ready.

Try the live validation demo, or request a walkthrough with your own data templates.

Try the live demoExplore the platform