DingTalk Goes Fully AI-Native with Version 8.0 Launch, Unveils Suite of AI Products and Hardware

On August 25, DingTalk held a press conference to launch version 8.0, marking the official transition of the collaborative office app with the largest user base in the mobile internet era to an AI-native platform. Wu Zhao, CEO of DingTalk, stated, “DingTalk 8.0 is also version 1.0 of AI DingTalk. We are determined to leave the past behind and, with a zero-based mindset, build a completely new DingTalk for the AI era.”

In 2015, DingTalk 1.0 was released. Over the past decade, DingTalk has defined the way we work in the era of cloud and mobile internet. DingTalk 8.0, codenamed “Fern,” is named after the earliest vascular plants on Earth, symbolizing the true maturation of life systems. It roots deeply into industries, reconstructs its core framework from within, and fosters a thriving ecosystem outwardly.

The all-new DingTalk has thoroughly disrupted and restructured the traditional form of office productivity tools, redefining how people interact, process information, search for knowledge, and create applications in the AI era. It has simultaneously launched over 10 AI-powered products, including DingTalk One, the enterprise AI search engine “AI Search,” AI Sheets, AI Notes, and the smart hardware device DingTalk A1. Starting today, users can download or update to the new DingTalk via its official website and major app stores to experience it immediately.

DingTalk One is the world’s first agent-driven workflow information stream, built by DingTalk for the AI era. It introduces a new paradigm of AI-powered information acquisition and collaboration, making work processing as seamless as scrolling through short videos.

In traditional work environments, the vast amount of work-related information we handle daily is scattered across multiple scenarios, including group chats, to-do lists, meetings, documents, and spreadsheets. This requires employees to proactively seek out information, which is not only inefficient but also hinders focus. DingTalk One leverages AI to organize all work-related information and tasks by priority, presenting them clearly to users via information stream cards, thereby shifting the paradigm from “people seeking tasks” to “tasks finding people.”

DingTalk One is also designed as a unified gateway for human-AI interaction through natural language conversations. Users can long-press the One button to initiate voice dialogues, directing various Agents within DingTalk to handle work tasks—such as scheduling meetings, creating to-do items, searching for information, organizing project progress, answering knowledge-based queries, or aggregating data from AI-powered spreadsheets—thereby significantly enhancing information processing efficiency.

Meanwhile, DingTalk One also serves as an internal corporate learning hub. It replaces the traditional model of disseminating corporate information primarily through group chats. Leveraging AI, enterprises can curate, translate, and push industry-related new products, technologies, trends, and insights to employees in a format akin to an internal corporate video channel or Douyin (TikTok).

AI search powered by large language models (LLMs) is gradually becoming the mainstream way for people to access information. However, general-purpose LLMs are often difficult to apply in enterprise scenarios. Work-related chat logs, documents accumulated by enterprises over the long term, meeting minutes from projects, and emails and approval workflows generated through collaboration are scattered across various platforms, making them hard to locate when needed.

DingTalk has officially launched its new enterprise AI search engine, “AI Seek” (AI Souwen), which integrates AI-powered search with AI-driven knowledge Q&A to provide employees with a modern approach to knowledge acquisition in the AI era. Whether searching for people, events, or knowledge, or inquiring about project progress, AI Seek delivers precise answers based on the user’s permission scope. For instance, when a sales department employee searches for “Q3 2023 sales revenue in East China,” AI Seek automatically aggregates relevant information within the user’s access rights—including progress updates from chat conversations, accessible spreadsheets, documents, daily reports, and data from the enterprise CRM—to generate organized sales data charts.

Furthermore, AI Search supports flexible integration of large language models (LLMs) and has innovatively launched the Model Fusion Engine, “AI Fusion,” which aggregates over 50 mainstream LLMs from around the world. Enterprises can select models within AI Fusion to compare question-answering performance, thereby identifying the most suitable model. To support Chinese companies expanding overseas, the platform also offers access to international models, ensuring compliance with local policies and regulations. For state-owned enterprises, financial institutions, and technology firms that prioritize data compliance, AI Search further supports the integration of on-premises, private, and industry-specific proprietary models, addressing corporate concerns regarding model security and operational efficacy.

Currently, more than 300,000 enterprises are building business applications using AI Tables on DingTalk. The new DingTalk has upgraded AI Tables into an Application Creation Platform for the AI Era (AI Infrastructure of Application), introducing a fresh visual design and new interaction paradigms, while launching multiple AI capabilities such as the AI Table Assistant and Field Agent.

With the AI Spreadsheet Assistant, users can automatically generate AI spreadsheets, automated workflows, and dashboards simply by describing their ideas in conversation. They can also directly perform data queries, formula-based calculations, or basic operations such as sorting, color-coding, and column widening, further lowering the barrier to using AI spreadsheets.

DingTalk AI Tables has launched over 100 field-specific Agents, equivalent to having AI assistants for more than 100 job roles automatically working within your spreadsheets. These agents cover multimodal AI capabilities such as AI-generated text, video, audio, and images, information extraction, image recognition, and speech-to-text conversion, as well as scenario-specific AI features like one-click virtual try-on for clothing images, one-click product modeling, and one-click background replacement.

In the safety inspection scenario of Shandong Weiqiao Pioneering Group, after inspection workers submit photos, the Field Agent can identify photo information one by one. AI then completes multi-stage analysis from hazard description to rectification suggestions. Rectification notices are pushed to department heads and responsible personnel through automated workflows, directly reducing accident rates.

Furthermore, DingTalk AI Tables, in collaboration with the Alibaba Cloud Yaochi Database team, leverages Alibaba Cloud’s technical expertise in data storage and computing to introduce O-Table, a novel integrated storage-and-compute architecture. This architecture delivers ultra-large-scale storage, AI-driven integrated storage and computing, and high-performance parallel processing, effectively addressing real-world scenarios involving massive datasets, frequent updates, cross-table associations, and real-time computation. Compared to similar solutions, it offers superior cost-effectiveness and practical usability.

DingTalk 8.0 Officially Launches the New Generation AI-Powered Voice Tool “AI Minutes” and Releases the Smart Hardware Device DingTalk A1.

DingTalk AI Listen & Note is a voice recording tool that leverages large language models to enable real-time speech-to-text transcription, semantic analysis, and intelligent summarization. Users can quickly launch the feature and automatically generate meeting minutes via video conferences, by searching for “Listen & Note” in the search bar, or by using DingTalk A1. The tool comes with 36 built-in scenario templates, comprehensively adapting to various occasions such as team meetings, class notes, interviews, psychological counseling, and legal consultations. It also supports enterprise-customized templates to meet personalized needs.

DingTalk A1 is the first AI hardware launched by DingTalk, working in synergy with its AI audio recording software and hardware. Compared to similar products on the market, the DingTalk A1 measures only 3.8mm thick—equivalent to the thickness of a few credit cards—and supports magnetic attachment to the back of a smartphone. Equipped with a six-microphone array and bone conduction technology, it achieves ultra-long-range audio pickup up to 8 meters along with high-definition sound quality. The device offers a battery life of up to 45 hours and provides users with 1,000 minutes of free speech-to-text transcription time and 10GB of cloud storage space.

DingTalk A1 deeply integrates AI-powered audio transcription capabilities at the software level, seamlessly connecting with DingTalk’s To-Do and Calendar applications. Additionally, DingTalk A1 can activate more than 10 specialized DingTalk AI Assistant personas, supporting user-defined configurations.

According to Wu Zhao, as of August 22, 2025, the number of enterprise organizations on the platform exceeded 26 million, including over 190,000 paying organizations; among the 5,191 companies listed on China’s A-share market, 79% have adopted DingTalk.

Currently, the DingTalk platform processes over 12 billion messages per day on average, with AI invocation volumes increasing threefold year-on-year, covering more than 20 industry scenarios including manufacturing, retail, finance, and healthcare.

In the healthcare sector, DingTalk has developed an efficient collaborative office solution for smart hospitals, addressing their modern workplace needs. It assists medical staff in automatically scheduling work hours and streamlining attendance management; supports high-definition video conferencing to facilitate remote teaching and academic exchange; and integrates with electronic medical record (EMR) systems to enable paperless operations, allowing physicians to easily access patient histories and formulate treatment plans.

Meanwhile, DingTalk has also partnered with the “Doukou Gynecology Large Model” developed by Yisheng Jiankang (Hangzhou) Life Technology Co., Ltd. This model has now passed the national written examination for senior professional titles in obstetrics and gynecology, marking the first vertical medical large model trained by a startup on DingTalk’s enterprise-exclusive AI platform to meet the standards of professional clinical examinations.

Beyond the aforementioned initiatives, DingTalk holds significant potential for further innovation in the healthcare sector. As Wu Zhao stated, “DingTalk will return to its entrepreneurial roots, re-embrace innovation, and reaffirm its user-first philosophy to build a collaborative office product worthy of the AI era.” As DingTalk’s AI capabilities mature, we can expect to witness more innovations from DingTalk in the healthcare field, empowering all hospital operational scenarios with efficient workplace logic.