Andeon turns your core banking, CRM, policy, credit-risk, and claims exports into a real RDF knowledge graph and a real Iceberg lakehouse — not copies of each other, the same source files, traced back to one build manifest. Ask a question in plain English, run SPARQL, or just write SQL.
Below is the actual home screen every analyst sees after logging in — not a mockup of one.
Your operational analytics platform powered by knowledge graphs. Select a module to begin.
Ask questions in natural language and get answers from the knowledge graph
Visualize entity relationships across all connected systems
Browse raw tabular data and RDF triples with filters
Query the same raw data as real Apache Iceberg tables (World B, ADR 0003)
Preview what Andeon would build for a new industry or use case — catalog, lineage, no activation yet
AI-powered analyst that reads the graph and proposes actions
Trigger operational actions across source systems
Unified view of a customer across all data sources
Four failure modes risk and data teams describe on the first call — and the exact module in the screen above built to close each one.
Core banking, CRM, policy admin, credit risk, and claims each hold a different slice of the same customer — reconciling them means opening five tabs and hoping the IDs line up.
Every source system is mapped into one OWL ontology and served as real RDF triples from Apache Jena Fuseki. Relationships between entities are one click away, not stitched by hand.
The knowledge graph is only useful if analysts can actually query it — and query languages aren’t part of a credit analyst’s job description.
Ask a question in plain English. An LLM translates it into SPARQL, runs it against the live graph, and shows you the generated query — nothing is a black box.
Most platforms export data into a second system for SQL analysis, and the two versions quietly diverge within a quarter — nobody trusts which one is current.
The same raw files are also registered as real Apache Iceberg tables, fronted by a lightweight Polaris catalog, queried live with DuckDB SQL. Same bytes, two query surfaces.
AS/400 has the account. Salesforce has the relationship. Policy Admin has the coverage. Credit Risk has the score. Claims has the history.
Search one ID and get a single unified identity, pulled live from all five systems, with every field labeled by the source it came from.
Every RDF graph and every Iceberg table trace back to the exact same raw file and the same build manifest. Lineage isn’t a bolted-on tracking product here — it’s just how the two worlds were built.
graph:CoreBanking, served live from Apache Jena Fuseki. Queryable by SPARQL or plain English.
iceberg.accounts, registered in Apache Polaris. Queryable live with DuckDB SQL.
No proprietary black boxes. Andeon assembles established open-source infrastructure and deliberately tunes it to run the whole stack on a single cloud instance per tenant.
Every environment ships via Terraform and cloud-init in minutes, at its own subdomain. personal.andeon.co is one of them — a real tenant, right now