Innovative Qizhi ChatBi's all-round upgrade helps factories explore digital intelligence transformation

The generative enterprise private domain data analysis application (ChatBI), which is based on the Kizhi Kong Ming AinOGC industrial model recently released by Kizhi Kizhi, is in the implementation process. With years of experience in implementing manufacturing scenarios and AIGC technology, Innovation Qizhi took the lead in testing ChatBI in a food and beverage manufacturer and new material factory, helping the production side achieve natural language interactive data analysis, making it easy for factory managers and business personnel to become data engineers, and helping improve production efficiency and quality.

Data analysis plays a critical role in actual factory production. How to enable business personnel to have timely and effective insight into information, discover and solve problems from information, thereby improving enterprise management efficiency and level of refinement is currently the core work content of manufacturing enterprises after digital information construction.

In the actual scenario of enterprise production, data analysis requires various departments such as operation, production, business, and technology to maintain the unification and alignment of requirements. This will incur huge communication costs. Not only is it time-consuming, but it is also easy to create gaps due to insufficient communication. On the other hand, in actual production, the scale of data must evolve as the business evolves, and reports need to maintain a high iteration frequency. Traditional dynamic data reports make it difficult for traditional dynamic data reports to respond quickly to changes in enterprise needs.

Innovation Qizhi's product ChatBi, based on the AinOGC industry model developed in-house, combined with the MI (Manufacturing Data Analysis System) of its subsidiary Oliver Chizhi (Manufacturing Data Analysis System), solves the technical problems of interactive intelligent query of production data and forms visual analysis reports for customers in one stop, and provides enterprises with intelligent data analysis and decision support to optimize production processes, improve production efficiency and quality.

On the premise of meeting factory business needs, ChatBI also fully considers the actual situation of users and provides a more convenient operation experience. For example, it provides data source management operations such as database table resource mapping and table data preview for the factory data terminal. Through simple settings, factory engineers can quickly access various OLTP/OLAP data sources such as MySQL, Snowflake, and Clickhouse, and view data in real time, making data access to data sources more convenient and visual.

A large-scale industrial model product that can be implemented in production scenarios requires users to flexibly search for data, and must also meet the needs of factory managers for data traceability. ChatBI not only finds the required tabular information in a huge database and generates SQL code based on the user's questions. After the SQL execution is completed, it automatically selects the most appropriate chart to present the data, and also generates data explanations to answer the user's questions. At the same time, it also supports the function of querying SQL code with one click, so that managers can feel more at ease with the generated data.

In order to be closer to actual business scenarios, ChatBI is based on the AINOGC industry model's multi-round dialogue technology, and has excellent ability to understand intentions and contexts. Become an exclusive intelligent assistant that knows more about corporate business, and can repeatedly help factory operators and managers explore more problems and freely explore more details and information.

By thoroughly understanding the business and data of the enterprise through dialogue, ChatBI can finally help enterprise managers quickly discover hidden problems and find optimal solutions from them, so that users can make decisions more quickly and intelligently, thereby helping enterprises improve efficiency and reduce risk.

ChatBI has been recognized by clients. First, it benefits from its ability to accurately transform users' natural language questions into effective SQL query statements. Its generation ability has been tested by various authoritative public data sets such as Spider and Chase-C, and its accuracy is superior to most mainstream large-scale model products.

ChatBi is able to quickly carry out tests at customer sites, thanks to the industry accumulation and knowledge recall technology that has been deeply involved in the manufacturing industry for many years with innovative wisdom. Help models understand the context of the problem when querying by proactively recalling industry knowledge. This enables ChatBI to provide users with the most professional answers according to the user's intentions, combined with industry expertise in the manufacturing field.

Ultimately, what can guarantee that ChatBi is accepted by manufacturing companies is a complete data security and privacy guarantee mechanism. Data security for manufacturing enterprises is essential. ChatBI supports the privatization and deployment of knowledge bases and large models in enterprises, and project delivery is ultimately used on the enterprise intranet to prevent the use of private enterprise knowledge for general model training. At the same time, it also supports data encryption within the enterprise according to different permissions, so that users can only query data within the authority, and desensitize or de-identify sensitive data to reduce data-related risks.