Ekiya applied in Tax law

Problem

Fiscal law is a complex matter with various sources of documentation: the law, drafted by the state’s representatives, it’s translation in a fiscal code and the implementation guidelines written by the administration’s experts and, finally, the case law resulting from court decisions.

All these documents evolve over time as new law is passed or new court decisions and this adds a complexity since the applicable rules depend on the exact date of the financial operations.

Solution

Ekiya indexes all information resources, keeping record of the origin of each bit of information. Ekiya is developing temporal and conceptual models that allow to contextualize each of the textual sources. The sources themselves are tagged in order to maintain the correct precedence (e.g. law over interpretations).

When a user asks a question, Wikiya finds the relevant documents, make a summary and points back to the original source for human control.

Benefit

Time saved: no more digging during hours to get to the right information.

Less litigations thanks to safer answers, based on the latest facts.

AI Technologies used

Ekiya´s Domain Knowledge Extractor (DKE) will firstly build a mind map from the documentation. This is done using Knowledge Graphs, which is visual representation of the relationships between different entities, facts and ideas. Each token, sentence, paragraphs is connected to elements of the knowledge graph.

Once the documentation is captured, the AI focuses on the temporal constraints and annotates each information as much possible with decision dates, applicability dates, etc.