AnalyticDB for PostgreSQL Exclusive AI Solution for DingTalk

Solution overview

AnalyticDB for PostgreSQL and DingTalk partner to offer an exclusive enterprise AI solution. AnalyticDB for PostgreSQL provides storage, vectorization, vector retrieval, and recall services for your private enterprise data. DingTalk provides the AI application solutions. This solution supports various scenarios, such as generating reports from natural language, searching documents, onboarding new employees, providing sales guidance, implementing intelligent customer service, and enabling smart product selection.

The AI application solutions include:

• ChatBI: Uses Natural Language to SQL (NL2SQL) technology to enable businesses to generate reports from natural language queries. It facilitates data delivery and collaboration through DingTalk IM and a PC web interface.

• ChatMemo: Provides a secure, scalable, and easy-to-maintain exclusive enterprise knowledge base. It offers capabilities like conversational knowledge retrieval, content summarization, and content generation. Through DingTalk IM dynamic cards, users can quickly search for and preview documents within group chats, making knowledge instantly accessible and boosting the productivity of managers and employees.

Additionally, the solution includes an exclusive AI management console for centralized management of your corpus and data sources.

Target audience

Businesses that use DingTalk for collaboration can purchase AnalyticDB for PostgreSQL and the exclusive AI solution for DingTalk to enable the ChatBI and ChatMemo solutions.

Key benefits

• You store your private enterprise data in a dedicated instance of AnalyticDB for PostgreSQL, which is exclusively yours. AnalyticDB for PostgreSQL further enhances data security with features like cloud disk encryption, SSL connections, and row/column-level access control.

• Unlike traditional reporting tools, this solution provides fine-grained access control based on your DingTalk address book. You can deliver data directly to individuals or groups through DingTalk IM dynamic cards, enabling seamless collaboration.

• A professional team provides expert model tuning services, and a self-service tuning platform is also available. This ensures that you can achieve high accuracy from your model outputs without needing in-house expertise.

ChatBI

The ChatBI solution enables users to generate reports by querying data with natural language, powered by NL2SQL technology. It delivers these insights and supports collaboration through both DingTalk IM and a PC web interface. The following diagram illustrates the technical architecture of ChatBI.

ChatBI lets you query operational data anytime, anywhere, using natural language—what you think is what you see. The following figure shows a real-world example.

ChatMemo

ChatMemo provides your business with a secure, scalable, and easy-to-maintain exclusive enterprise knowledge base. It lets users retrieve knowledge through conversation, summarize content, and generate text. The following diagram illustrates the technical architecture of ChatMemo.

The following video shows a use case where a logistics company integrates with a DingTalk bot to display content using dynamic cards in the DingTalk client.

Exclusive AI management console

AnalyticDB for PostgreSQL provides an exclusive AI management console for each enterprise. This console allows you to manage your corpus, optimize models, and upload and maintain local files.

The ChatBI solution, which is oriented toward data assets, is available on a graphical engineering platform, allowing you to deploy it yourself.

The exclusive AI management console provides three configuration tabs: Vocabulary, Experience, and Sample. These tabs are used to manage business terms and their associated items, configure query experience rules such as result limits and format specifications, and define sample question-and-answer pairs with their corresponding SQL statements, respectively. In the Modify Application dialog box, you can set the Application Type (ChatBI, ChatMemo, or ChatForm), select the Model (such as GPT16K), and configure the maximum number of turns for conversational memory. In the Set Data Source dialog box, you can select the Data Source Type (such as MySQL, GaussDB, PostgreSQL, or API) and configure the Connection URL, Username, and Password.

More information

For more details about the exclusive AI solution for DingTalk, please join the DingTalk group with the ID 59370008826.