Botfactory, the hottest DIY one-stop enterprise

2022-08-07
  • Detail

DIY's BOT factory, a one-stop enterprise emotion AI platform

on May 29, 2020, Zhujian intelligent held a bot factory 2020 new product launch conference with the theme of AI smart and simple new infrastructure, launched a newly upgraded one-stop enterprise level emotion artificial intelligence platform, and launched a detailed demonstration of the platform design concept and specific function highlights around "dialogue AI", "cognitive AI" and "emotional AI" starting from the current tide of new infrastructure

Zhujian released its first version of BOT factory in 2017. After four years of optimization, it has completed the upgrade iteration of multiple versions, which has become the key core and pillar for the commercialization of Zhujian products

this release includes 27 important technology and function upgrades and 6 enterprise level intelligent application releases. One platform, n scenes, no code, fast landing, designed to create a dialogue robot platform that everyone can create

01 challenges in enterprise intelligence transformation and providing customers with information opportunities

new infrastructure is centered on the new development concept. Under the catalysis of new infrastructure, artificial intelligence is the first to accelerate the implementation in enterprises. Intelligent upgrading has become the key to enterprise transformation

what intelligence needs to solve is the core problem of enterprise development: cost reduction and efficiency increase. The greatest value of dialogue robot is to provide new ideas for problems such as limited human efficiency, high human cost and high customer acquisition cost, so as to realize multi role, full scene and full media collaborative application

many enterprises have adopted dialogue robots, but new problems have arisen in the process of using them:

operation and maintenance are becoming more and more difficult

business scenarios cannot be updated

answers are irrelevant, menu based, and unable to answer customer questions

knowledge is not connected to business data

lack of effective self-learning

for the problems of enterprises in scenario use, It is a thankless thing for insurance companies to launch the one-stop housing endowment service. The enterprise level emotional artificial intelligence platform BOT factory helps create a management dialogue robot. The robot is controllable, manageable, attainable, able to understand intention and emotion, and supports 0 code development and training, deployment and maintenance. It has a smooth experience of the whole process and scene. The enterprise does not need to configure professional technicians, and everyone can start it

02 three AI pillars build a complete platform

bot factory 2020 integrates dialogue AI, cognitive AI and emotional AI

conversational AI refers to the use of artificial intelligence technology to process and realize conversational interaction. The realization of dialogue interaction requires a strong semantic understanding ability, the basic ability of NLP, as well as voice, knowledge map, recommendation and other technologies. Conversational AI requires a large number of AI basic module integration to support it. From algorithm, capability to platform, it supports developers to carry out secondary development, and enables everyone to participate in the use, update and maintenance

cognitive AI includes technologies such as knowledge atlas, which can process a large amount of unstructured data, such as product documents, regulatory documents, financial statements, etc. at the same time, it can also continuously learn from the experience of data mining, NLP and intelligent automation, so as to help enterprises process a large amount of data

emotional AI is reflected in the emotional intelligence of dialogue robot, that is, dialogue fluency. Emotional ai not only understands people's emotions, but also ensures that dialogue and interaction are like people. Zhujian provides 25 kinds of emotion recognition. In addition, it also includes knowledge reasoning, rich skills, accurate intention, multi scene jump, full duplex voice, dialogue behavior detection, user portrait and other technologies. In addition, it has accumulated a large number of dialogue scenes and data to ensure the double quotient of dialogue robots

build a one-stop enterprise artificial intelligence platform through dialogue AI, cognitive AI, emotional AI, ability + data + scenario, on which scenarios and solutions can be built

03 the four major technology bases enrich the platform scenario

the core technology base, including the technologies accumulated by Zhujian through years of self-study. It mainly focuses on the six technology cores of natural language processing (NLP), knowledge engineering, in-depth learning automl, text data platform, speech recognition and visual computing, and takes into account a number of AI technologies such as voice, image and semantics to provide the ability to understand, understand and read. In terms of core NLP capabilities, in some specific areas, models for training document recognition and document parsing are added to improve the accuracy of knowledge extraction. In addition, in terms of the advantages and architecture of Chinese natural language understanding technology, Zhujian intelligence has upgraded the English semantic model

the second layer is the AI platform layer, which consists of two major platforms, namely, BOT factory and Gemini. BOT factory platform is a platform for full scene and end-to-end development of interactive dialogue robots. It can develop TOC robots or tob businesses. Gemini is a platform built around knowledge innovation and application. It helps enterprises use AI mining, automatic construction and other technologies to process massive unstructured data and semi-structured data into knowledge maps, and develop various intelligent applications in combination with business characteristics

the third layer is the application layer. Various intelligent applications developed based on the platform can directly complete the scene landing. It includes the AI contact center total solution based on speech semantics and the application based on knowledge map. These application functions have been verified in practice. They can integrate the existing systems of the enterprise and quickly upgrade to intelligent business systems. In the future, we will also work with our partners to release more enterprise level intelligent applications

the fourth layer is a complete industry-based solution, which accumulates into a standard industry solution through intelligent application in different industry scenarios. In the application of smart finance industry, AI technology is used for precision marketing, customer activation, intelligent investment adviser, financial analysis, fund map, financial business handling, financial collection return visit, crowd portrait analysis, etc

04 everyone can DIY robot platform

bot factory 2020 has three characteristics: pre training intention to easily create personal assistant, enterprise assistant and business assistant; Hundreds of built-in skills can be used immediately after opening the box; Chat scenarios are built quickly

0 code creates a robot. After it goes online, the robot continuously improves the answer accuracy through self-learning, and has the characteristics of simple maintenance. After upgrading the English semantic model, you can create English robots on BOT factory 2020, and even make outbound calls in English

drive dialogue scenarios with business objectives to build iterative dialogue scenarios

the NLP method is used to solve the error correction difficulties such as inaccurate transcribing and missing words in speech scenes. At the same time, it has upgraded many understanding technologies, such as dialogue behavior classification, dialogue information extraction, accurate intention recognition, etc. The fine-grained model is optimized, including business context, long sentence multi intention understanding and emotion recognition in different fields

the scenario is connected with the business data, and the knowledge is automatically updated

it covers six industries: finance, intelligent terminal, medical and pharmaceutical, intelligent education, Internet and traditional industries. The domain modules are started quickly to realize the self-learning of robots

05 three drives and three deployments

bot factory 2020 has made important upgrades to the three main engines and achieved the integrated user experience of QA + knowledge + task

qa driven

solves the problem of cold start. Through historical log data learning + industry pre training model, the creation data cost of dialogue robot is very low. Through the batch test of dialogue, the accuracy of question and answer of dialogue robot after online is guaranteed

knowledge driven

by decoupling data and semantics, knowledge entities can be added continuously without affecting the robot's understanding of problems

unstructured documents can be transformed into knowledge maps through automatic processes of identification, analysis, extraction and construction, and can be searched and reasoned

task driven

man machine dialogue should adhere to the principle of continuity and smoothness. The bamboo dialogue robot not only supports the continuity of context within a single task, but also supports the continuity of context between task scenarios

task robots cannot extract simple keywords, so the dialogue robot among bamboos solves the technical difficulties of classification, extraction, recommendation and so on semantically. Without writing code, you can build task scenarios and seamlessly integrate voice through drag and drop workflow

after iteration and project verification, the tools of the task engine not only ensure a friendly interface for setting up scenarios, but also support development level code integration, taking into account the use experience of ordinary operators and professional developers. All accesses can be accessed in the form of API. Friendly and rich APIs support ecological development

flexible deployment mode ensures that enterprises can choose according to their own needs. The BOT F1 pull machine is mainly supported by the electromechanical industry 2020 for public cloud, private cloud and hybrid deployment. The advantage of the public cloud is that it does not need hardware cost, opens on demand, and ensures low latency and high concurrency dialog requests, which is suitable for small and medium-sized enterprises

private cloud is suitable for large enterprises. It adopts high availability architecture, supports qpsxn performance expansion, and ensures security and monitoring, including virtualization

06 jointly build an AI ecosystem

in addition to new products and new upgrades, Zhujian intelligence also launched the AI partner plan in this press conference. By recruiting AI acting partners, AI regional partners and AI strategic partners, a cooperative ecosystem of bamboo intelligence is established. Zhujian will provide business, training, service and marketing support for future AI partners, so that more people will become beneficiaries of new infrastructure in the 21st century

Zhujian intelligence adheres to the development of naturallanguageprocessing technology as the core, constantly polishing from product ecology to business ecology. It is said that NLP is difficult to be a unicorn. Indeed, this is a race track to compete for technology and endurance. Only by persisting can we usher in the bright future of dialogue robots

Copyright © 2011 JIN SHI