In the wave of generative AI technology, the industry is facing a series of deep thinking beyond technology: When applying AI, do we really understand the role of AI in reshaping social business? In this Microsoft Envision AI Connection Shanghai event, through the practice sharing of AI pioneers, the pain points, challenges and experiences faced by enterprises in the process of AI implementation were revealed, and the importance of thinking transformation and scenario development was emphasized.
The first quarter of this year has passed, and a new wave has been set off in the field of generative AI, which has evolved from a broad technical concept to a segmented and vertical application. For example, the recently released Microsoft Copilot for Security (international version) has become the first independent generative AI solution in the global information security business field to date, and there are more segmented and vertical AI applications, which are effectively changing the working model of related vertical industries.
However, a series of deeper questions are gradually emerging: have we stopped to think about the rapid change of the technology frenzy: will blindly catching up with the new trend inevitably bring about positive industrial change, and do enterprises really understand the profound impact of AI technology before rushing to put it into practice, and how it will reshape our society and business ecology?
At the just-concluded Microsoft Envision AI Connection Microsoft AI Innovation Forum Shanghai Station, many AI pioneers stood at a key turning point and traced a general path for enterprises to implement AI - in order to successfully land AI and truly release the role that large models should play, two key problems need to be solved first.
First question:
Before AI transforms, do we need to transform our thinking first?
On the way to the implementation of AI, what is the pragmatic idea?
"More than 200 years ago, engines were invented and were very much like today's AI, which would be revolutionary, but some coachmen were thinking about how to put the engine in a carriage. Wei Qing, CTO of Microsoft (China), shared such a thought experiment from the Massachusetts Institute of Technology on the development path of AI - "putting the engine on the carriage" may be the fastest application method that people can think of, but it must not be the best and most representative of the trend. Each generation of technology has its own unique mission, and Al's more important mission is to create new models rather than simply and linearly perpetuating traditional thinking. "We should design new cars, plan new roads, and even create a whole new era based on new engines, rather than continuing traditional thinking," Wei Qing said.
Therefore, the so-called pragmatism is not about blindly pursuing speed and "new trends", but rather focusing on how to implement AI, how to build innovative productivity that adapts to the new era, and finally build the transformation ability to enter the next era.
The source of this pragmatic thinking first requires an in-depth analysis of the real challenges and pain points faced by enterprises when applying AI and large language models.
Victor Lee, head of the Digital Innovation Center at Carl Zeiss (Shanghai) Management Co., Ltd., mentioned that although the company has invested a lot of resources in AI, they are still exploring and verifying the most basic question - "where can AI help employees and business", rather than simply using AI as a tool to improve efficiency. Hu Chengchen, chief expert and assistant vice president of NIO's technology planning, believes that the biggest challenge for NIO in applying AI at present is how to grasp the two sides of technology, avoid achieving it overnight, and gradually implement it according to its own actual situation, and make use of its strengths and avoid weaknesses. Ren Feng, Co-CEO and Chief Scientist of Insilico Medicine, shared that the application of AI is solving the pain points in the life sciences and pharmaceutical fields, but domestic companies are generally still unaware of the importance of AI, resulting in hesitation in decision-making.
These practical sharing reveal a common problem for the industry: the biggest difficulties and pain points on the road to AI implementation are often not in technology, but in the ideology of managers and decision-makers. Whether it is too impatient to pursue AI applications or too conservative to ignore its potential, it is a common problem in the current process of AI implementation.
As Wei Qing emphasized, the implementation of enterprise AI is a set of system engineering methods, in which talents, data, algorithms, and computing power are all indispensable, but the most important thing is the transformation of talent capability models and thinking in the AI era, which changes the status of AI from "AI for Data" to "Data for AI", and finally realizes "Tech for human". In fact, it is only a shift in the mindset of people that can transform AI from a tool on the periphery to a center of change.
In this regard, Tao Ran, Chief Operating Officer of Microsoft Greater China, concluded: "The implementation of AI is never just a technical issue, but an ideological issue. "Generative AI, as a new way of production, cannot be seen as just a tool. At this critical moment of technological paradigm shift, Microsoft's most worthwhile and most important role to play is "Making AI Real for Enterprise" - to better link large models with enterprise business scenarios.
Therefore, in the face of a new round of progress in generative AI technology, enterprises need to think and respond with a more pragmatic attitude, not only focusing on the rapid development of technology, but also deeply thinking about how to truly integrate these technologies into enterprise operations and innovation, and create a new era.
Second question:
With the technology, where are the use scenarios?
In the process of implementing AI, although the ideological level needs to be carefully considered, there are also many things that are worth trying quickly and accumulating experience. McDonald's China's rapid response in the field of AI in the past year provides a case worthy of analysis and reference for the industry.
Tang Haitao, Vice President of Digital at McDonald's China, believes that McDonald's, as a leading global food service company, needs to seize new technologies and market opportunities through AI and innovate on this basis. In the context of digital transformation, AI empowerment provides new development opportunities for enterprise practices: first, to optimize employees' work experience, second, to improve production efficiency, and third, to integrate AI technology in the field of business development to continuously explore new opportunities for enterprises.
Through the cooperation with Microsoft, McDonald's China has promoted the application of innovative technologies in various fields such as AI-assisted operations, employee empowerment, skills training, and technology research and development, and has achieved unexpected results and positive user feedback.
At the forum, more Microsoft partners shared their experience from exploration around "practical AI landing scenarios". Hu Chengchen of NIO proposed three methodologies of "identifying AI landing scenarios", combining "AI capabilities" with "business pain points", "own advantages" and "user demand insights", realizing the layout and innovative application of AI in all fields of NIO, including human-vehicle interaction, intelligent manufacturing, service system, internal project management and other fields, and innovatively combining generative AI technology with NIO's strengths in cloud and computing power, further strengthening NIO's advantages over its competitors.
Insilico Medicine Ren Feng shared an exciting real-world story. In the development of drugs for the treatment of idiopathic pulmonary fibrosis, Insilico Medicine has leveraged generative AI technology to unleash the advantages of AI in screening a large range of targets and rapidly generating compounds in new drug development, shortening the early development cycle of four and a half years to about 18 months, and reducing the investment of tens of millions of dollars required for conventional research and development to $2.6 million.
In addition to accelerating innovation in the innovation track, generative AI technology is also helping enterprises transform and upgrade in some traditional industries and open up new growth curves. Carl Zeiss, a German company with a history of 178 years, has used AI to do things that could not be done before, such as using computer vision to inspect optical products and improve accuracy, and has also used AI to "unleash" many things that do not need to be done. Carl Zeiss Victor Lee shares: "We use AI to optimize processes to find out which parts of everyone's daily work are unnecessary and remove them. As a century-old company, Carl Zeiss has been able to welcome the AI era with ease.
From the above successful cases of AI pioneers, it is not difficult to see that regardless of the scale of the industry and enterprise, the implementation of AI needs to explore application scenarios according to the actual situation and achieve differentiated and special applications. These scenarios are not fixed and vary from enterprise to enterprise, but it is precisely because of this difference and particularity that the scenario-based application of generative AI technology is also directly related to the comprehensive management and even the transformation of business thinking of each enterprise. As Lou Xuejian, General Manager of Microsoft Technology Center, concluded, the widespread application of AI is an irreversible development trend of the times, and enterprises should embrace AI technology with a more open attitude to drive enterprises to discover more possibilities and a better future brought by the use of AI.
Tao Ran, Chief Operating Officer of Microsoft Greater China, also said that Microsoft will always insist on turning AI technologies and capabilities into reality in enterprise scenarios, and strive to promote AI to achieve more commercial value in enterprises. It is precisely because of this development orientation that Tao Ran continued to share the professional Copilot (international version) application launched by Microsoft based on multiple vertical business scenarios such as employee office productivity, customer service, sales, IT professionals, and security analysis, helping users greatly improve work efficiency in various work scenarios. For example, based on user feedback data, Copilot for Microsoft 365 (International Edition) helps users save 85% of their time in drafting new documents and documents quickly.
Yin Jing, CTO of Microsoft Greater China Industry Solutions, further analyzed the practical path of enterprise construction of generative AI technology in a more pragmatic and detailed manner, including four main stages: GPT casing, developing interactive applications with external capabilities, developing AI native applications, and building agents.
There are more application scenarios for generative AI technology, which can be seen not only in the services provided directly by Microsoft, but also integrated into the innovative business of many ecosystem business customers. At the Microsoft Envision AI Connection Shanghai AI Innovation Forum, companies such as Shuosoft, Avanade, Hanshow Technology, and Nanyang Wanbang cooperated with Microsoft to showcase a variety of rich, vertical, and pain point solving services, covering enterprise-level generative AI development, enterprise operation management, smart retail, and enterprise knowledge base, and a new technology ecosystem is taking shape. As Wei Qing emphasized in the opening sharing, compared with technology, Microsoft hopes to provide enterprises with new ideas and methods, and jointly open a new era of productivity.
When exploring the characteristics that enterprises need to have to adapt to the next era of AI productivity, it is not difficult to find that in addition to maintaining open innovation, pragmatic thinking, and accurately grasping practical application scenarios, there are more key elements worthy of in-depth thinking and preparation in advance, and similar forward-looking insights and deep insights can be seen everywhere in this event. For example, Yin Jing also emphasized that the start-up teams of some AI startups can achieve high market valuations with only a dozen employees, precisely because the computing power they can mobilize is enough to rival that of a traditional enterprise with thousands of employees.
These trends and insights will also continue to inspire Microsoft and its partners to think about the essence of technology, grasp the laws of application, and work with partners to overcome transformation obstacles, better embark on the AI trend, and pragmatically solve AI transformation problems at the dawn of the new era.
Follow @MicrosoftTechnology to learn more about the industry