On the afternoon of January 9, at the "Super Assistant" session of DingTalk 7.5 product launch conference, Qi Junsheng, chief product officer of DingTalk, gave an in-depth explanation of DingTalk's thinking and practice on intelligent products. He said that DingTalk We are committed to lowering the threshold for using AI so that everyone can create AI assistants, allowing every individual and enterprise to expand their capabilities and gain a more efficient and intelligent work experience.
Comprehensive embrace of AI, new features of DingTalk AI Assistant
Since the full launch of intelligence in April 2023, DingTalk has changed rapidly. Qi Junsheng pointed out that with the continuous advancement of AI technology, especially the continuous improvement of large models, the progress on the capability side is very obvious. Great progress has been made both in terms of model performance and cost efficiency. At the same time, the demand for intelligence in various industries is becoming increasingly clear, which provides DingTalk with more opportunities and challenges.
In order to better meet customer needs, DingTalk continues to iterate its products, from the announcement in April that DingTalk would join Tongyi to launch intelligent upgrades, to the comprehensive scenario reconstruction and transformation in August, and to this time As of now, DingTalk has released AI assistants for 20 product lines and more than 80 scenarios that have been fully integrated into large models. The service scope of AI assistants covers almost all DingTalk product lines. Qi Junsheng said that this is a reflection of DingTalk’s professional capabilities, pragmatic attitude and entrepreneurial spirit to truly implement AI practices.
As an important milestone in DingTalk’s advancement, AI super assistant is of great significance in the current context. From the perspective of product thinking, Qi Junsheng summarized the four new features that support AI super assistant products as new capabilities, new interactions, new scenarios and new creations.
In terms of new capabilities, the AI super assistant has the ability to interact with external information in real time and accurately interpret it. This is due to the rapid progress of large models and AI PaaS, as well as the large amount of valuable real and effective data that DingTalk has accumulated over the past few years. These data are ultimately deeply integrated with model capabilities, underlying capabilities, and platform capabilities.
In terms of new interactions, the LUI GUI model is reducing complexity into simplicity and reconstructing the entire product experience of DingTalk. "The AI assistant is always on call and at your fingertips. Just ask @ and a processing task will appear. This should be an agile and capable assistant around you." Qi Junsheng explained that with the support of the AI assistant, new interactions will be recognized The user's current state, such as a meeting or document scene, through multi-modal methods, AI assistants can better understand the user's intentions and needs, thereby providing more personalized and efficient interactive services.
In addition, in terms of new scenarios, the AI super assistant will continue to expand and deepen its application scenarios. Qi Junsheng mentioned that customers' needs have been very clear. They hope to integrate project data, human resources data, sales and marketing data, etc. to form comprehensive decision support. AI super assistant meets this demand and seamlessly connects with existing systems to provide enterprises with a smarter work experience. Previously, Belle had provided store managers with real-time sales data analysis through the "Smart Questions" service, further releasing the management energy of "store managers"; and at the Social Organization Service Center of the Sichuan Provincial Department of Civil Affairs, the " "Intelligent Q&A" provides policy consultation and service guidance to various organizations, and has won recognition and praise from the local service public.
Finally, there is a new creation. "Whether enterprises and users can truly create their own AI assistants in a very efficient and low-threshold way is what DingTalk values most." Qi Junsheng said, whether it is personalization, local learning, native AI calling, or The connection and integration of external systems and applications can be easily "digested" by the AI assistant. This is also DingTalk’s expectation, so that everyone who has unique skills and professional knowledge and is willing to do more can gain greater energy through AI assistants.
Build differentiated AI platform capabilities and accelerate the move towards "AI"
Qi Junsheng emphasized that a wide range of customer needs and capability supply are the core differences of DingTalk's AIAgent. DingTalk has rich scenarios and data accumulated by the company itself, which will enhance the Agent's capabilities.
He emphasized that DingTalk AI Assistant is not only the in-depth thinking and practical implementation of AI Agent by DingTalk, but also the crystallization of joint wisdom with customers. During the product development process, DingTalk has an in-depth understanding of the needs and pain points of enterprises in various fields. Based on four "new features" that represent differentiated application scenario values, the AI assistant has developed unique platform capabilities.
Qi Junsheng introduced that the AI platform capabilities include scene perception, memory enhancement, action fusion and base blessing. Specifically, perception determines whether the AI assistant can accurately understand user intentions; memory is crucial to the performance of AI products, and DingTalk will have a great advantage with the support of such data and true identity decision-making. The rich application link capabilities and underlying capability blessings will make the AI assistant easier to call, more open, and stable to release and take on the user's AI imagination.
The arrival of the AI era brings unlimited room for growth. Thinking about the future, Qi Junsheng said that moving from AI to AI is a very clear trend. In the future, more new AI and AI-native product forms will appear, and APPs will also move towards the Agent model.
To welcome the future, DingTalk will also provide multi-level and multi-form capability supply to individuals and ecological partners, create personalized and rich AI assistants, promote the final form of the AI assistant market, and jointly create a new round of business formats, ecology, business model. During this process, "remain in awe of the times and products," Qi Junsheng added, "a low-threshold, open, and high-frequency AI assistant platform is the goal that DingTalk has been pursuing."
The above is the detailed content of DingTalk Qi Junsheng: Rich scenarios and enterprise needs are the biggest differences in DingTalk becoming an AI agent. For more information, please follow other related articles on the PHP Chinese website!