After the one-year anniversary of ChatGPT's release, what is the next high ground for large models?2023 is regarded as the first year of large-scale models, and 2024 is regarded as the first year of large-scale implementation of large-scale models.
In November 2023, Bill Gates published an article in which he said that AI (agent) will be the next platform after the big model, not only changing the way everyone interacts with computers, but also completely changing our lives within five years.
Nowadays, more and more large-scale model companies and technology companies have begun to deploy agents. For example, Kunlun Wanwei released the "Tiangong Skyagents" platform, and Zhipu AI released the Cogagent model, which has vision-based GUI agent capabilities. Feishu, a subsidiary of ByteDance, has launched an AI product "Feishu Smart Partner", which can be regarded as a combination form of agent. OpenAI's upcoming GPT Store is the AI Agent Store that it has created.
In 2023, large model companies will be busy finding computing power and refining models, and AI application layer companies will be busy finding scenarios for AI transformation or AI native applications. This round of generative AI has been hyped as the industrial revolution in the AI era, so it needs to meet not only the entertainment needs of ToC such as writing poems and paintings, but also the needs of improving productivity in the ToB field. In 2023, large models will mainly stay in the AI "toy stage" of the former, and this year they will enter more AI "tool stage".
Therefore, in 2024, AI agent will be the next flashpoint after the large model. If the large model is the infrastructure of water, electricity and coal in the future, then the agent is the way for users to contact and use AI in the future. For many businesses, it's time to ask yourself, "Are you AI ready?"”
Companies that are the first to achieve AI ready have taken the lead in embracing AI. Feishu is one of the most representative tools.
Anker Innovations is a global consumer electronics company, and one of the most important things to do is consumer insights. They collect user feedback from the sales process, classify and tag it, and store it in their own QMS quality management system. In the past, if you want to use these data to guide the business in reverse, such as figuring out why consumers return goods and which channel they often return, you need to equip a special R&D team, log in to the system, improve data, and make reports.
Last year, through Feishu's intelligent partners, Anker Innovations is carrying out a new round of iterative upgrading of the QMS system, and product managers, R&D teams, and customer service teams can all use the Q&A method to ask questions directly on Feishu to quickly understand the channels, reasons, quantities, and next improvement measures for returns. William, R&D Engineer at Anker Innovations, said that Feishu's smart partner has turned the quality management system into a "quality think tank", "like snapping a finger", and these key insights are all around and are really put to use.
Feishu is not the only company that can do AI, but its unique feature is that Feishu plays the role of not only a digital tool, but also an AI co-creator. Feishu is positioned as a comprehensive work platform with "ail in one". The overall product concept of the latest Feishu 7 is "business at hand", and on the basis of fully precipitating enterprise information, the interface is open to third-party systems. For enterprises that solve the digitalization of vertical scenarios, Feishu's "all in one" product positioning is just complementary to it.
Xinheyun is a manufacturing SaaS manufacturer that provides "MES+ERP" factory management systems for manufacturing customers, and its digital terminals have chosen to cooperate with Feishu. Tao Binjiang, the founder of Xinheyun, once shared his thoughts on cooperating with Feishu: "As a vertical SaaS vendor, Xinheyun's core competitiveness is to solve the business problems of the manufacturing industry, including production, research, supply, marketing, and service (production, R&D, **chain, sales, and service), but it does not need to develop a terminal entrance by itself. This is an important reason for choosing Feishu as the entrance to the digital terminal. ”
At present, a Feishu is about equal to "Slack + Zoom + Google Doc + Workday + Airtable + Teambition", and it may continue to be added in the future, all integrated on one platform. This makes Feishu an important conduit for enterprises to achieve digital transformation naturally.
On the other hand, while helping enterprises achieve digitalization, Feishu intelligent partners can naturally undertake the results of digitalization and help enterprises enable AI with one click. Feishu's smart partners have knowledge, memory, initiative, and can also go deep into the business. In business scenarios such as content creation, content summarization, data analysis, scenario construction, and system construction, users can work together with intelligent partners. In addition, as an open AI service framework, Feishu Smart Partner allows enterprises to independently select the appropriate underlying large model according to their business scenarios.
According to public information, Feishu Intelligent Partners has carried out in-depth co-creation with several companies, in addition to the aforementioned Anker Innovations, and also jointly explored the possibility of AI efficiency improvement in store management scenarios with Yuanqi ForestTogether with enterprise service company Jimmi Technology, Feishu Smart Partner reshapes the sales sceneCooperate with e-commerce marketing company Chase Media to create an efficient business process ...... with AI
Xinmin Evening News reporter Jin Zhigang.