ARCHITECTURE OF A VALIDATION ORCHESTRATOR IN A MULTI-AGENT INTELLIGENT COMPETITIVE INTELLIGENCE PLATFORM

Abstract

The article addresses the problem of automated validation of bibliographic and factual sources in the context of multi-agent intelligent competitive intelligence platforms. The relevance of the study is driven by the growing volume of information flows and the need to ensure the reliability of data used for management decision-making in organizational systems. The aim of the work is to develop the architecture of a validation orchestrator that provides hybrid routing of requests between APIs of scientific databases and web resource parsing modules. A multi-level orchestrator architecture is proposed, including a routing layer, a unified evidence schema, a cross-verification mechanism with a four-level reliability scale (Confirmed, Likely, Unknown, Conflict), and a request audit subsystem. A comparative analysis of API-based and parsing-based approaches is conducted against criteria of stability, source coverage, reproducibility, and completeness of extracted data. The scientific novelty lies in substantiating a hybrid approach to validation with a formalized evidence aggregation schema from heterogeneous sources.