A society entering the digital era poses challenges of industrial restructuring to businesses. Digitalization typifies dynamic information, whose uncertainty promotes enterprises to get into deep understanding of the characteristics and approaches of digital transformation. At the AI-Empowered Digital Transformation High-end Forum and FTEMBA program admission briefing meeting, Wenxuan Ding, Professor of Artificial Intelligence and Business Analytics and Deputy Director of Business Intelligence Center (BIC) of emlyon business school, presented a comprehensive analysis on how AI helps enterprises to carry out digital transformation and clarified the important role of AI in such a process.

The essence of enterprise transformation, as pointed out by Professor Ding in her speech, is to replace the decaying productive forces with the new and advanced ones, and to re-explore and re-construct the industrial ecology and production organizations that adapt to the advanced productive forces, which is not possible without the progress of science and technology and the improvement of the cognitive and technical capacity of all the members of the enterprise.

Digital transformation emphasizes the approaches of transformation – the “how” aspect: AI is currently the most matured and publicly available high-end technology, whose work foundation and processing content are digital numbers.

AI can be divided into weak AI and strong AI based on whether it considers and imitates the cognitive and thinking processes of the human brain. Weak AI is characterized by its reliance on big data but it does not have the ability of understanding, thus is suitable for pattern recognition, automated complicated calculation, target movement, chess playing, etc. Strong AI does not necessarily depend on big data but is capable of understanding and possesses self-awareness.

To give software a function, said Prof Ding, is to plant in the machine a willpower to realize the function.

Then, how does AI enable a business to transform? Professor Ding gave the answer from two perspectives: the upgrading and the reconstruction of industrial structure.

The upgrading of industrial structure means to broaden and deepen the industrial restructuring through the innovation of products and services without changing the industrial foundation. AI can provide three types of empowering:

1. Empowering through optimization. AI is used to accurately process data of various representation relationships without changing the original operations.
2. Empowering through expansion. The existing business domains are broadened and deepened while AI is used to analyze the market trends. New products and process will be created according to the current competitive advantages. Examples include the addition of homogeneous products (e.g., the chips developed by GREE) and the development of heterogeneous products (e.g., the greenhouse vegetables cultivated by Tencent)
3. Empowering through process changes. AI is used to revolutionize business models without changing the existing products (e.g., the success of PDD -- a platform for group buying deals)

The industrial restructuring requires the corporate leaders to have an insight into the changes of markets, revolutionize the existing industrial structure, find the sunrise industries and achieve an overall innovation. In this case, weak AI can be used to assist the design of new products, and strong AI for making independent innovation and scientific discovery.

Generally speaking, the essence of AI empowering is to reconstruct the enterprise with AI as an energy source.

Through AI, enterprises can realize the digital transformation in many aspects: productivity (production means change from energy conversion tools to intelligent tools), production materials, production methods, organizations of production, producers (consumers can also be producers), products, locations for production, etc.

Therefore, enterprises need to better understand the design of AI algorithms and software, which is similar to the design of an anthropomorphic vitality. At present, AI generally relies on big data, with sensors, Internet of Things, and statistical machine learning as the lifeline. A reliable and accountable AI should meet the following five requirements:

• Able to capture the real-time evolution process of the state;
• Useful data, not necessarily big data;
• Focusing on "causal" relationships rather than "correlations";
• Avoiding a "black box" operation and being interpretable; 5. Human-centered.

When it comes to the digital transformation, a traditional industry should take into consideration the following questions: What enterprises must be transformed? Where to go and what is the targeted task? How and what to transform? Who is going to "do" this "transformation"? How to convert the work mode and service content of "management"? How to change the "labor (employees)"? If you want to succeed, what must you do, change, learn, invest in and give up immediately? How much fund do you need and where does it come from? How to deal with old production plants, products and businesses? Who will create new products and businesses and when will they be effective? How long is the transition period from "old" to "new", how to get through it smoothly, and how to deal with the change? What are the loss estimates for the transition period? How do investors change? What methods do shareholders use to supervise digitalized and intelligent enterprises?

Digital transformation can turn an enterprise into an optimized physical entity supported by advanced productivity, capable of dealing with uncertainties, reducing risks, creating value and enhancing competitiveness. In the future, when quantum computers and biological computers are fully developed, the enterprises will be elevated to a higher level of productivity.