News / All News / Article

A Look at the Core Capabilities of an Intelligent Data System

Article images may contain Chinese text.


AI-Driven Conversational Analytics, Paired with a Unified Data Foundation



The first core capability is AI-driven conversational analytics paired with a unified data foundation.

The system supports ingestion of multiple data formats from multiple sources—different databases, imported files, and other data types can all be brought together.


At its heart is the AI agent's conversational interface. Just as you would with the AI tools you use every day, you ask a question directly, such as "What is the hospital listing status of a particular product in a particular region?" and it immediately provides results.


It also supports multi-turn conversations. If a chart is not quite right, just say "Switch to a line chart" or "Switch to a pie chart," and it adjusts in real time.


You can also @mention entities to invoke configured knowledge graphs and business rules, putting the entity relationships you have built up directly to work for more accurate, faster queries. Not satisfied with the first answer? No problem. Just as when chatting with an AI agent, keep asking follow-up questions and making corrections to move closer to the result you really want. Data exploration is no longer a one-shot gamble: your questions make the results increasingly precise.


Ontology-Driven Semantic Layer Construction



The second core capability is ontology-driven semantic layer construction.

The system uses a BRAG-driven relationship traversal approach to transform traditional two-dimensional tables into a networked data structure.


Take a retail outlet, for example. Its associated metrics, responsible staff, target attainment rate, and market share are connected organically into a semantic knowledge network, rather than scattered across isolated wide tables. Compared with traditional two-dimensional table configurations, this approach can improve data-query accuracy by 30% to more than 35%.


How do you configure it? When you import a customer data table containing target attainment, metrics, gaps, profiles, and more, the system automatically identifies the relationships between entities—building the structures, attributes, and links for you.


You can also make adjustments and additions manually in the interface. That is the core advantage of an ontology-driven approach and knowledge graphs: data is no longer a set of isolated tables, but a network of relationships and semantics.