Directly hit WAIC丨Accenture: More than half of Chinese companies are still in the experimental stage of artificial intelligence applications

21st Century Business Herald reporter Yang Qingqing and intern Shi Jie reported from Beijing

Every company aspires to be a "reinventor" under the AI wave, and technology will play a key role.

The birth of ChatGPT has stimulated an unprecedented wave of AI, and more and more enterprises are actively exploring and deploying artificial intelligence to seek business innovation and industrial upgrading. Today, generative AI and big language models are rewriting industries. According to Accenture, 42% of global companies surveyed intend to invest heavily in the generative AI and big language models behind ChatGPT this year.

Recently, during the 2023 World Artificial Intelligence Conference, Accenture held a thematic forum on "Generative AI: Reshaping the Enterprise", and interpreted the previously released "Technology Outlook 2023" report, trying to explore how enterprises should stand on the cusp of the tide and achieve high-quality growth under the new wave of technology.

Hong Zhu, Global Vice President and Chairman of Greater China, Accenture, said at the forum: "In the next decade of digital transformation, enterprises must be completely reshaped, and artificial intelligence will become a key driving force for enterprises to comprehensively reshape the era. ”

At the same time, as more and more Chinese companies explore and apply artificial intelligence, the problem is becoming more prominent. "Most Chinese companies are still in the experimental stage of applying AI." Yu Yi, president of Accenture's Greater China Enterprise Technology Innovation Group, said frankly in an interview with 21st Century Business Herald.

"More than half of Chinese companies lack AI-savvy technical personnel and high-quality datasets. At the same time, problems such as unclear AI application scenarios and high cost of investing in AI projects have also brought certain problems to Chinese enterprises. Yu Yi said, "The key to solving is that enterprises make good use of existing resources and make personalized adjustments according to their own needs." ”

Reinvent your business

The advent of the big model has become one of the most significant leapfrog changes in the history of artificial intelligence evolution, which brings challenges but also means opportunities, which no enterprise can ignore.

According to Accenture, after the birth of ChatGPT, 98% of executives believe that generative AI will be able to reshape the entire operation of the enterprise, and 40% of man-hours in all walks of life will be assisted and replaced by artificial intelligence technology based on large-language models.

"We believe that in the next decade of digital transformation, enterprises must be completely reshaped, and artificial intelligence will become a key driving force for enterprises to comprehensively reshape the era." Zhu Hong said in the forum.

Yu Yi also said: "The new functions of the basic model and the continuous advancement of technology are regarded by some people in the industry as an important turning point towards strong artificial intelligence. While only time will tell if the technology and methods behind the underlying model will be sufficient in the future to enable some form of truly strong intelligence, the degree to which the underlying model is versatile in certain data types is still significant and could completely change the way and scenario in which enterprises apply AI. ”

Yu Yi further pointed out that in the future, generative artificial intelligence and big language models will empower and innovate the following work scenarios.

First, consultation and advice. Typical areas include customer support, sales enablement, human resources, corporate strategy and market intelligence; The second is content creation. Generative AI can reveal new ways to reach and engage audiences, while bringing speed and innovation in areas such as production design, design research, visual identity, name development, copy generation and testing, and real-time personalization.

The third is to write code. Software coders will dramatically increase productivity with generative AI, quickly translate one programming language to another, master a variety of programming tools and methods, automate code writing, and anticipate and prevent problems; Fourth, automation. In the future, some business process automation will be pushed to new and transformative levels;

Fifth, security protection. Over time, generative AI will support organizations in strengthening governance and information security, preventing fraud, improving regulatory compliance, and proactively identifying risks by establishing cross-domain connections and inference capabilities both inside and outside the organization.

"It can be expected that future generative AI could have a huge impact on science, enterprise data, product design and manufacturing." Yu Yi said.

In addition, Yu Yi also pointed out: "We cannot look at the application of artificial intelligence in isolation. As companies begin to enter the 'total reinvention' phase, technology adoption should not be single-point, siloed. To make it clear that the overall progress of the enterprise is a systematic and holistic matter, it is necessary to consider the whole business from the perspective of the business. ”

AI application pilot phase

The expected impact of large models on businesses cannot be underestimated. At present, more and more Chinese enterprises are actively exploring artificial intelligence technology to seek efficiency improvement and business innovation.

In a recent Accenture Technology Trends Survey, 72% of 225 Chinese executives surveyed were very or extremely excited about the new capabilities enabled by AI big models.

According to the Technology Outlook 2023 report, there are currently two different ways for enterprises to apply generative AI and big language models.

One is to use applications, that is, directly purchase "model as a service" to conduct business applications. At this stage, some large technology companies and research institutions have completed the construction of pre-trained basic models to use them as platforms to support new AI applications. It provides basic model services in the form of open source channels or API paid access. Therefore, downstream enterprises do not need to build their own basic models, and can focus on relying on existing models to enrich content.

The second is to customize the model according to your own needs. Although the base model has been pre-trained and has strong adaptability, it may require more targeted training for some downstream tasks. So companies can use their own data to customize or fine-tune models to meet unique needs.

For the two application methods, Yu Yi said, "Most companies need customized models to expand their use and value. ”

In his view, custom models allow companies to fine-tune preset large models with their own data, enabling them to support specific downstream tasks across the business. At the same time, by understanding the company's historical background and business operations, customized models will also help enterprises effectively enhance employee capabilities, improve customer satisfaction, introduce new business models, and timely perceive upcoming changes.

However, while Chinese enterprises are chasing the wave of AI, they also encounter difficulties and challenges.

"Most Chinese companies are still in the experimental stage of applying AI." Yu Yi said frankly. "More than half of Chinese companies lack AI-savvy technical personnel and have not accumulated high-quality data sets. In addition, problems such as unclear AI application scenarios and high investment costs in AI projects have also brought certain troubles to Chinese enterprises. ”

In his view, many difficulties encountered by enterprises in the application of artificial intelligence technology are mainly caused by two reasons. First, companies lack a clear AI roadmap to move AI projects from proof-of-concept to actual production. Second, enterprises cannot fully understand the overall development pattern of artificial intelligence, which also leads to enterprises easily following the inherent behavior model, that is, starting from scratch and building cars behind closed doors.

In this regard, Yu Yi said: "The key to solving is that enterprises make good use of existing resources and make personalized adjustments according to their own needs. ”

In addition, the report also points out that enterprises need to rethink and re-examine their AI strategies, fully consider the best application scenarios of large models in enterprises, and better grasp the new opportunities brought by AI.

For more information, please download 21 Finance APP