Common meaning thousand questions, first "ask" DingTalk

At the 2023 DingTalk Ecological Conference, a “digital clone” that can replace real people participating in the conference will appear, along with “AI assistants” who help teachers correct homework and AI recruiters from companies. These “sapient” colleagues use DingTalk as an entrance to enter the real work scene.

Digital employees are an example of DingTalk’s “reengineering” of products with the help of Ali Tongyi Qianwen large model. Next, DingTalk will open up AI PaaS capabilities and replace its ecological applications with AI soil.

Two years ago, DingTalk began to transform from a SaaS application to a PaaS platform. As of the end of March this year, DingTalk had more than 5,000 ecological partners. Now, it will use large models to help these applications complete intelligent reengineering faster.

In DingTalk’s AI application market, 10 transformed application products have been launched, covering business scenarios such as intelligent personnel and intelligent education. Compared with the general-purpose large models of major companies such as Baidu and Ali, which are adjusted and transformed by various industries, DingTalk, which has its own application ecology and vertical business, appears to be lighter in pace, and the scene is closer to the “efficiency improvement” that AIGC focuses on. .

The pace of using AI large models to transform mobile Internet applications is getting faster and faster.

Reshape DingTalk with Tongyi Thousand Questions

Digital employees and even digital clones have really appeared. They are not digital people who rely on people in the middle to work, but “sapient people” who have IQ and can understand human speech.

At the 2023 Dingding Ecological Conference held on August 22, Ye Jun, President of Dingding, demonstrated the working scenarios of multiple digital employees.

In audio and video conferences, employees can use “digital clones” to attend the meeting instead of themselves to help them understand the important information of the meeting, and push the summary of the meeting after the meeting; for the education industry, “AI small teaching assistant” can help teachers intelligently correct homework, accumulate Learning data, generate lesson preparation suggestions and PPT for explanation, etc.

In terms of corporate recruitment, DingTalk and No. 1 Zhipin co-created a digital employee “Recruitment Specialist (No. 1 Recruitment)”, which can help HR write recruitment notices, post them to corresponding recruitment websites, collect resumes, screen resumes and make appointments for interviews.

Even, AI employees can help colleagues order and put together orders. The “Happy Fight” function jointly created by DingTalk and Ele.me allows users to order and put together orders in natural language in the DingTalk group without leaving the group. You can pay with one click after chatting, and the internal applications of the Ali system are linked again in the form of AI.

DingTalk brings AI digital employees

Digital employees in different positions can enter the company’s address book, obtain the corresponding job authority like human employees, and participate in the company’s organizational governance and business collaboration. The main job of these digital professionals is to liberate human colleagues from boring and repetitive routine work, so that they can focus more on creative work.

The emergence of the aforementioned AI employees is a phased result of the intelligent transformation of DingTalk.

On April 18 this year, Dingding announced that it will be connected to Alibaba’s Tongyi Qianwen large-scale model, and stated that it will use the large model to redo Dingding. At that time, DingTalk released an AI “magic wand” that demonstrated intelligent applications in four high-frequency scenarios, including chat, documents, and audio and video conferences.

Four months later, DingTalk handed over a new answer sheet at the Ecological Conference. Ye Jun said that DingTalk has 17 product lines and 55 scenes fully connected to the large model to achieve intelligent transformation.

More importantly, the access of AI has opened up more room for business growth for DingTalk.

According to the charging standard announced at the ecological conference, on the basis of the annual fee of 9,800 yuan for the professional version of DingTalk, corporate users can get 200,000 large-scale call quotas by adding 10,000 yuan; 450,000 large model calls. According to this calculation, the cost of calling the large model is less than 5 cents each time.

There are applications that are easier to AI

The focus of this ecological conference is not only the intelligent upgrade of DingTalk itself. What is even more eye-catching in the industry is that DingTalk has opened an intelligent base AI PaaS (Platform as a Service) to ecological partners and customers. With the help of this underlying AI platform, SaaS (Software as a Service) applications in the DingTalk ecosystem can quickly access AI ability to complete intelligent reengineering.

DingTalk Opens AI PaaS Capabilities

Moving from an enterprise service-oriented SaaS application to a PaaS platform has been the focus of DingTalk’s efforts in the past two years.

In 2021, DingTalk came up with the idea of “doing ecology”. It started with a low-code (aPaaS) solution, allowing developers to quickly build applications for enterprises through DingTalk’s low-code platform, even for non-technical personnel. Create simple applications by “drag, drop, drop” using ready-made tools. Since then, DingTalk has successively launched bPaaS (cool application), iPaaS (connection platform), dPaaS (data platform), etc., to provide a convenient development environment for ecological developers.

At the ecological conference in March last year, DingTalk once again emphasized the importance of ecology to it. It proposed the “PaaS First Partner First” strategy, expressing its determination to transform to a PaaS platform.

After two years of development, DingTalk’s application “Castle” was first built. Ye Jun introduced that as of March 31 this year, DingTalk had more than 5,000 ecological partners, including independent software developers (ISVs), consulting ecosystems, sales and delivery service providers, and hardware ecosystem manufacturers. The number of DingTalk applications exceeds 10 million, of which the number of low-code applications exceeds 8 million, which paves the way for the subsequent ecological AI.

Today, DingTalk has decided to transfer AI capabilities to applications in the ecosystem after transforming itself with the large model of Tongyi Qianwen. AI PaaS acts as the “soil” for intelligence.

This large-scale intelligent transformation of applications has already begun. At present, DingTalk has launched the “AI Mall”. In this AI intelligent application market, there are already 10 intelligent SaaS applications.

Take the “i HR system” as an example. This application can automatically check whether the content of the labor contract complies with laws and regulations, and evaluate the risks of the labor contract; at the same time, it can answer questions about labor and personnel, and provide suggestions, real cases, and sources of evidence etc.; it can also automatically identify and evaluate labor dispute risks, brain drain risks and employee satisfaction risks.

Another AI application “Easy Whale Cloud” is an AI low-code platform. AI can automatically generate an application outline that meets the needs based on chat content or a given application name and detailed description. Users can continuously adjust the corresponding application modules by talking with AI until the desired application is polished.

These AI applications have provided a sample for the intelligent upgrade of DingTalk’s entire ecosystem. They will also continue to infiltrate the actual office scene and change the traditional way of working from the user side. In Ye Jun’s words, the large model capability is really used for work, turning it from a “toy” into a productivity tool.

Compared with major companies such as Baidu, Huawei, and Ali who use the smart cloud to export large-scale model capabilities to various industries, DingTalk has its own application ecology and is vertical to the office scene, and the pace of promoting application intelligence is brisk and logical.

Ali’s Tongyi Qianwen first blossomed on DingTalk. Looking at it from a larger perspective, these “flower branches” that have just broken ground will form a joint force with other AI applications on the market to jointly build an ecological environment to inspire and give birth to more intelligent applications.

DingTalk’s practice of reborn with AI also shows that the ecology with applications is easier to AI.