Enterprise-AI

Product Introduction Product name: Ai Assistant Enterprise AI Platform Product positioning: Based on Azure AI services, it supports enterprise privatization deployment, low-cost implementation, and solves the "last mile" problem of enterprise AI application landing. Product advantages: （1） Provide enterprise level privatization deployment solutions, build enterprise specific AI applications based on enterprise private data, and ensure data security and privacy. （2） Rely on Microsoft Azure AI Service to achieve compliance call of various large models, self-developed LLMOps engine layer (including knowledge enhancement engine, knowledge retrieval engine, agent orchestration engine, and multi model gateway) supports integrated application development, deployment, optimization of multiple scenarios, as well as enterprise knowledge insight and data insight, and accelerates enterprise intelligence transformation. （3） Out of the box: personal AI portal for enabling enterprise employees to work, knowledge management and application for enabling enterprise transformation and development, to meet the knowledge question and answer scenario of L1-L4. （4） Enterprise level management and control meet the requirements of necessary management and control for enterprise privatization to promote AI applications. Product module: Module name Function Overview Personal assistant The personal efficiency tool created for enterprise employees enables employees to work in office scenes and improves work efficiency. Its functions include: document interpretation, document translation, poster generation, PPT generation, in-depth reasoning (exclusive reasoning model), online search, prompt word square, etc. Enterprise Knowledge Center 1、 Establish a closed loop of enterprise knowledge from construction, application, renewal and precipitation. 2、 Intelligent and accurate knowledge retrieval, personalized knowledge services. 3、 3. Support multiple modes such as cross database questioning and selective database questioning. 4、 Support the warehousing (including internal and external data sources) and output of multimodal knowledge. 5、 The knowledge base supports documents in various formats, online web pages, and enhanced parsing of images and scanned PDFs. 6、 Support users to configure "knowledge expert agent" by themselves and seamlessly integrate with various systems, and carry out various internal and external application scenarios. 7、 QA collection, manual feedback and knowledge compensation mechanism, knowledge leakage prevention, support for refined authorization of knowledge base, and controllable knowledge opening. Agent Center 1. Based on agent orchestration engine and visual configuration tool, make AI agent with zero code; 2. Based on the Microsoft AutoGen framework, it provides low code standard agent development, which can realize process based agent scenarios: For example: intelligent customer service, contract review, document proofreading, resume scoring, work order assistant, marketing assistant, report generation, etc; 3. Custom agents can be exported as JSON configuration files, seamlessly integrated into any Python application, or encapsulated in Docker containers for separate deployment on various platforms. MxNetAI engine 1、 Knowledge enhancement engine 2、 Knowledge retrieval engine 3、 Agent orchestration engine 4、 Multi model gateway Enterprise level control 1、 The hot plug mode can flexibly connect various large models, small models and tools. 2、 Graphical presentation of multiple dimensions: number of conversations, token consumption statistics. You can view token consumption, conversation interaction times and other data from the perspective of applications, models, users, and departments. 3、 Flexible department, user and role permission management based on the enterprise organizational structure. 4、 Keep and view user logs and conversation records. 5、 Comprehensive system configuration: open API configuration, unified identity authentication configuration, interface logo, application parameter settings, etc.