JitWord协同AI文档团队
发自 JitWord AI 旗下AI产品：JitWord协同AI文档，JitKnow AI知识库，Pxcharts超级表格
最近一直在迭代 JitWord 协同AI文档 和 JitKnow AI知识库。 在大量用户的反馈和建议下，我们今天发布了 4.7.0 版本，致力于打造更懂业务的AI知识库系统。
我们都知道，企业的知识资产，分散在文档、表格、制度、方案等各个角落，跨资料信息整合非常困难；另一方面 AI 问答有时会答非所问，业务智能体搭建门槛非常高，一直是企业落地 AI 知识库的痛点。
所以在新版本里，我们围绕知识推理、Agent 业务编排、技能复用三大方向深度升级，进一步释放企业知识资产价值，让知识真正驱动业务落地。
下面我就来详细和大家聊聊这次我们的功能更新。
✨功能更新复盘
1. 知识库检索回答能力增强｜支持多文档关联推理回答
传统知识库大多只能单文档片段检索，面对需要综合多份资料的复杂业务问题，回答容易碎片化。
JitKnow AI知识库的新版本中，强化了 RAG 检索增强能力，实现多文档关联推理回答。AI 可同时读取、关联多份知识库文档，综合不同文档里的条款、数据、流程信息，整合输出完整、逻辑连贯的答案，回答有据可溯源，大幅减少 AI 幻觉，复杂业务问题也能获得完整可信的解答，检索准确率进一步提升。
不管是查阅跨制度合规条款、综合多份项目资料做方案解读，还是整合多份产品手册解答业务疑问，无需人工翻阅多份文件，AI 自动完成跨文档分析推理，降低员工信息查找成本。
2. AI 工作流优化｜拖拽配置 Agent 客服助理，零代码搭建业务智能体
Agent 可视化工作流能力我们做了进一步优化，无需编写代码，通过画布拖拽即可快速配置 Agent 客服助理。
我们可以把知识库检索、多轮对话、意图识别、文档处理等能力节点自由串联，快速搭建适配企业自身业务的客服智能助手。
同时，对于内部员工问答、外部客户咨询、售前售后接待等场景，也可以快速落地，7×24 小时提供智能服务，降低人工接待压力。
业务人员也可以自主调整 Agent 逻辑，不用依赖开发团队，快速响应业务变化，业务流程自动化落地更简单高效。
3. 全新技能库管理｜能力沉淀，一键应用到 AI 问答
之前有客户和我们反馈是否支持 Skill 技能库，最近我们技术团队也立马支持了技能库管理模块，把高频业务能力、提示词、处理逻辑沉淀为可复用技能资产。
企业可以将业务问答模板、专项处理逻辑统一收纳到技能库；需要的时候，一键将技能挂载到 AI 问答助手，直接赋能知识库对话。 实现能力一次配置，多处复用。这样可以保证新业务场景不用从零搭建，直接调用已有技能，极大缩短 AI 应用搭建周期，团队优秀业务经验可以沉淀、流转，避免能力随人员流失。
同时我们还支持一键导入技能，如下图所示：
可以快速将企业已经沉淀的技能一键导入使用。
在AI问答助手模块，我们也添加了技能库的选项，可以一键应用指定的技能：
不止新功能，成熟企业级底座持续护航
JitKnow 作为企业级 AI 知识中台，原生支持 PDF、Word、Markdown 等 20 + 格式文档导入解析；支持多人协作编辑、细粒度权限管控、版本回溯；支持私有化部署、源码交付、OEM 定制，数据完全留存企业内部，满足金融、医疗等行业合规要求，配套完整开放 API，可无缝对接企业现有业务系统。
目前已经覆盖了企业知识库、智能客服、员工培训、合规管理、研发知识沉淀、销售赋能等多元业务场景，帮助企业把沉淀的文档资料，变成可问答、可推理、可自动化执行的知识生产力。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
JitWord with the AI Document Team
AI product from JitWord AI: JitWord with AI document, JitKnow AI knowledge base, Pxcharts Supersheet
Recently, JitWord has been working with AI documents and JitKnow AI knowledge base. With the feedback and advice of a large number of users, today we have publish version 4.7.0, dedicated to developing a more business-informed AI knowledge base system.
As we all know, the knowledge assets of an enterprise are scattered across documents, tables, systems, programmes, etc., and it is very difficult to integrate information across the board; on the other hand, AI questions and answers are sometimes asked, and the operational AI agent is set at a very high threshold and has been a pain in the way of the establishment of the AI knowledge base.
So, in the new version, we're further unleashing the value of corporate knowledge assets and making knowledge really drive business down.
I'm going to talk to you in detail about our new functionality.
✨Phone Update Disk
1. Knowledge base search response capabilities enhanced to support multi-document-related reasoning responses
Most of the traditional knowledge banks can only be retrieved in single document segments and are prone to fragmentation in the face of complex operational questions requiring the integration of multiple pieces of information.
In the new version of the JitKnow AI Knowledge Bank, RAG search enhancement is enhanced to achieve multi-document-related reasoning answers. AI can read at the same time, link multiple knowledge-base documents, synthesize articles, data, process information in different documents, integrate the output of complete and logical answers, and provide evidence-retroactive answers, significantly reduce AI hallucinations, and provide complete and credible answers to complex business questions and further improve retrieval accuracy.
Whether it is to access cross-system compliance provisions, integrate multiple project information for programme interpretation or integrate multiple product manuals to answer business queries, there is no need for manual review of multiple documents, AI automatically completes cross-document analysis reasoning and reduces the cost of finding staff information.
2. AI workflow Optimizing Drag and Drag Configuration Agent Passenger Services Assistant, AI agent
Agent visualizes the workflow capability, and we are fine-tuning it. We don't have to write codes, so we can quickly configure Agent's passenger service assistant by drag and drag.
We can freely associate the skills nodes of the knowledge base, multiple rounds of dialogue, intent recognition, document processing, and so on, and quickly build up the customer intelligence assistants of the appropriate companies' own business.
At the same time, internal employees can be asked questions and asked questions, external clients are consulted, and pre-sale and after-sales receptions can be provided quickly, with 7 x 24 hours of intelligent services to reduce manual stress.
Operators can also adjust the Agent logic independently of relying on development teams to respond quickly to business changes and to automate business processes more simply and efficiently.
3. New skills bank management capacity is sunk, and one key applies to AI questions and answers
Before that, there was feedback from clients and whether we supported the Skill Skills Bank, and recently, our technical team supported the Skills Pool Management module, which has sunk high-frequency operational capabilities, prompt, processing logic into reusable skills assets.
Businesses can integrate business question and answer templates and specialized processing logic into the skills pool; when required, they can load skills to AI question and answer assistants, directly enabling knowledge base dialogues. Achieving a single configuration of capabilities and multiple reuses. This will ensure that the new business landscape is not built on a piecemeal basis, directly deploys available skills, significantly shortens the AI application set-up cycle, and that good business experience of the team can sink, flow and avoid loss of competencies.
At the same time, we support one key import skills, as shown in the following figure:
You can quickly import the skills of an enterprise that has settled.
In the AI question-and-answer assistant module, we have added the skills pool option, which allows for the application of specified skills:
It's not just new. It's a mature enterprise base.
JitKnow, as the firm-level AI Knowledge Centre, supports the import of 20+ files in PDF, Word, Markdown, etc.; supports collaborative editing, fine permission controls, version retrieval; supports privatization deployment, source code delivery, OEM customisation, data retention within the enterprise, compliance with industry requirements such as finance, medical, etc., and complete access to the existing business systems of the enterprise.
Multiple business scenarios, such as the enterprise knowledge base, smart customer service, staff training, compliance management, research and development knowledge deposition, and marketing empowerment, have been covered to help enterprises turn the sedimentary documentation into a question-and-answer, reasoning and automated knowledge productivity.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.