How to make full use of McKinsey’s 10 insights into artificial intelligence?

There is no need to change the original meaning, the content that needs to be rewritten is: Image source: Visual China

Rewrite the content without changing the original meaning. The language needs to be rewritten into Chinese and the original sentence does not need to appear

After long calls, McKinsey & Company finally released the TOP10 2023Q3 Artificial Intelligence Insights Report, of which reports on generative AI accounted for half.

The report "The state of AI in 2023: Generative AI's breakout year" analyzes the deployment and use of artificial intelligence by enterprises in various regions and industries around the world, and believes that the widespread use of artificial intelligence will have an impact on all walks of life. Tremendous influence.

Next, we will review this report from China’s perspective

01 Global generative AI adoption: What makes leading companies strong

First, although generative AI has been public for a short time, people are not unfamiliar with it.

According to the survey results, 79% of the respondents said they had been exposed to AIGC, while 22% said they often used AIGC at work

At the enterprise level, AIGC is also widely used, with one-third of respondents saying that their enterprise already uses AIGC regularly in at least one function.

Among respondents who said their businesses use AI, 40% said their companies expected to invest more in AI as a result of AIGC, and 28% said, The use of AIGC is already on the agenda of corporate boards.

Second, high-performance AI companies are more inclined to use artificial intelligence in product and service development. Unlike most companies that use AIGC as a means of cost reduction, the primary goal of AIGC for high-performing AI companies is tocreate new business or revenue sources, and improve existing products based on new features of artificial intelligence the value of.

Respondents from these enterprises noted that the top challenges they face are issues related to models and tools, such as monitoring the performance of models during production and retraining models as needed. In contrast, other interviewees mentioned broader strategic issues

Regarding the differences between leading AI companies and other companies, Shi Dao once pointed out in the article: Leading AI companies have certain commonalities in five aspects: governance, deployment, partners, personnel and data availability. If you are interested, you can click to view.

Third, the demand for artificial intelligence-related talents has changed. Over the past year, the most common hires by companies adopting artificial intelligence were: data engineers, machine learning engineers, and artificial intelligence data scientists. Compared with previous surveys, a much smaller proportion of AI-related software engineers are now hiring (28% in the latest survey, down from 39% in the previous survey). In addition, with the widespread application of AIGC, the demand for prompt engineering skills among enterprises is also increasing.

Shidao believes that the artificial intelligence talent pool may also affect the speed of domestic enterprises deploying AI. Another McKinsey report pointed out that Chinese companies focus on recruiting technical talents such as software engineers and data engineers, but ignore key talents such as AI translators and designers. It is reported that in 2022, nearly half of Chinese companies will employ software engineers, while only 6% of companies will recruit translators, which is far lower than the 14% in leading countries. This data deserves attention.

Fourth, the report states: Although the use of AIGC tools is rapidly gaining popularity, the data does not show that these new tools are driving enterprises to fully adopt artificial intelligence. For now, the proportion of companies adopting artificial intelligence has generally remained stable, with 55% of respondents stating that their companies have adopted artificial intelligence.

However, most respondents reported an increase in AI-related revenue in every business function where AI is applied. Looking ahead, while overall AI adoption remains at around 55%, more than two-thirds expect their companies to increase investment in AI over the next three years.

02 China’s generative AI industry: still in its early stages

Generative AI has entered a period of rapid growth. At present, China's market size is about 1/10 of the global total, and is expected to reach 200 billion yuan by 2025, with an average annual growth rate of more than 60%.

There is no doubt that investors will enter the market one after another in the future, and potential enterprises and individual users have expectations for the functions of generative AI tools. These all indicate that the spring of the industry has arrived.

What are the characteristics of the current domestic GenAI (generative AI) industry?

Specialized Hardware: Despite limited infrastructure resources, the localization process continues to accelerate

Due to the inability to obtain high-end GPU chips A100 and H100, domestic large-scale model research and development institutions have encountered bottlenecks in computing power. Although China's local chip R&D manufacturers have strengthened scientific research and the computing power of newly developed chips can reach 2000TOPS (currently the computing power of domestic GPUs is mostly within 1000TOPS), there is still a certain gap between the 4000TOPS computing power of the internationally leading H100 chip.

Basic model: Industry large models may be the way out for domestic GenAI.

What needs to be rewritten is: on the one hand, although general large models need to be trained, the number of Chinese websites in the world only accounts for 1.4% of the total, and the number of public Chinese corpora available for training is limited and the quality is uneven. On the other hand, due to the protection of user privacy, most of the data generated by domestic users on websites and mobile applications cannot be used for training large models, which has an adverse impact on the training efficiency and accuracy of the model

Therefore, compared with general large models suitable for multiple fields and tasks, industry large models rely more on developers’ understanding of vertical scenarios and the support of massive industry data. In the context of limited computing power, industry large models are more likely to become a winning track for Chinese GenAI manufacturers.

Application: China’s startups are still in their early stages and mainly focus on some specific fields

Domestic startups in the GenAI field have advanced financing rounds, concentrated between angel rounds and Pre-A rounds. Most of the financing amounts are less than 100 million yuan. They are in the early stages of development and have huge industry potential.

In the overseas GenAI industry, the scale of scientific and technological innovation companies is relatively large, and many unique unicorn companies have emerged in niche application fields. Among them, a global AI research company headquartered in London is valued at US$3.8 billion

In addition, the domestic GenAI vertical application field mainly focuses on text, image and audio and video generation, while overseas markets have seen a large number of GenAI-based development platforms, data analysis platforms and code writing platforms outside the above fields. The reason is that there are many systems written in early programming languages ​​overseas, and many companies face high programming labor costs, so there is a high demand for programming assistance software.

03 The industry is concentrated and most companies are still exploring business models

The commercial application of China’s GenAI industry presents two major characteristics:

First, the industry distribution is concentrated, mainly in advantageous industries with relatively mature commercial development.

In China, the fastest growing fields of GenAI applications include e-commerce, media, entertainment and games, especially digital virtual humans and e-commerce video marketing, while most traditional industries (such as finance, energy, education, etc.) are still in the Small-scale pilot stage.

The reason why GenAI applications can flourish on Internet e-commerce platforms is that China has a high-quality e-commerce and supply chain ecosystem as well as a huge consumer group, which provides opportunities for the implementation of GenAI applications

Typical industry application cases include: a generative AI 3D short video content manufacturer in the video industry, a leading GenAI video large model R&D company in the e-commerce industry, and a game and AI research and application institution under an Internet platform in the game industry wait.

Among the above-mentioned companies, there are some generative AI 3D short video content manufacturers. They can generate 3D video content based on text and promote it through short video platforms such as Douyin, Kuaishou, and Bilibili. In this way, the creative efficiency of short video content producers has been greatly improved

GenAI video large model R&D company is deeply involved in the e-commerce industry, using AI to generate virtual human anchors to empower e-commerce companies in various industries and brands.

The Game and Artificial Intelligence Research and Application Institute has used artificial intelligence technology to develop AI companion and AI competitive robots, and used natural language processing technology to give non-player characters characters so that they can interact with players in the game. Dialogue

Second, most companies are in the early stages of actively exploring their own business models.

The industry distribution is too concentrated, and it is more oriented to C-end users. This is a problem. It sounds great, but there are actually some challenges that need to be faced

As Shidao has discussed: The Internet era is different from the big model era. In the Internet era, "traffic" is free. When the total operating cost remains unchanged, the more users there are, the greater the value of the network. In the era of large models, computing power has a cost. Every time an additional user is added, computing power must be paid in real terms, which makes subsidies meaningless. The more new users you have, the less money you may make.

Therefore, targeting B-side enterprise customers may better represent the direction of future artificial intelligence companies.

In the European and American markets, large-scale enterprise customer groups are the mainstream. The SaaS paid subscription model has basically matured. A number of overseas GenAI companies have taken advantage of the mature SaaS market to apply GenAI to high-tech, communications and various traditional industries (such as medical care, education, etc.), with a wider range of applications

In the domestic market, the target customer groups of GenAI companies are mostly terminals. User companies are not willing to pay for software, the market needs to be further cultivated, and companies have data security concerns about SaaS deployment methods. The business model for large-scale application of GenAI still needs to be explored.

At present, the mainstream business models in the domestic market include cloud resource sales, model API calls, software-as-a-service charging, material charging, etc.

In general, most GenAI startups in the domestic market have just completed the output of standardized products and have begun to enter the preliminary commercialization exploration stage. As China’s SaaS market matures and enterprises’ willingness to pay increases, domestic text generation and image generation start-ups are expected to rapidly expand into the enterprise customer market.

Six Soul Questions: How does artificial intelligence technology translate into economic benefits?

Research shows that only 9% of Chinese companies plan to achieve revenue growth of more than 10% by deploying AI, while 19% of companies in leading countries are expected to achieve this goal. Similarly, in terms of profit contribution rate, only 7% of Chinese companies said that AI's contribution to EBIT exceeded 20%, while 14% of companies in leading countries exceeded this ratio.

The results show that Chinese companies urgently need to improve their liquidity and convert the potential of AI technology into tangible economic benefits.

In addition to the above-mentioned insufficient reserve of AI translators, the reasons why Chinese companies lag behind leading countries in deploying AI may also be:

1. The overall AI strategy is poor: among the Chinese companies surveyed, less than 30% have AI strategies that are consistent with the company’s overall strategy; among the executives interviewed, only 25% fully recognize the AI ​​strategy.
2. Not paying attention to internal training: Only about 30% of Chinese companies rely on internal training to cultivate AI talents, which is significantly lower than the global average of 45%.

So, in order to deploy generative AI as soon as possible and successfully extract value from it, enterprises must first think about six key questions:

1. In which business links can generative AI be deployed to create the highest value? What are the key use cases that can help you increase your competitive advantage?
2. What are your most important data assets that can be learned by AI?
3. What is your technical operation model?
4. Do you have the AI ​​talent you need to seamlessly integrate business and technology and turn the potential of AI into value?
5. Have you developed a risk protocol to mitigate model risks in generative AI (such as dealing with model “hallucinations”)?
6. How do you plan to promote change management to promote generative AI and achieve business goals?

McKinsey pointed out that companies must answer the above questions and overcome various challenges so that they can quickly build the capabilities required to fully unleash the potential of generative artificial intelligence and capture the economic benefits that this disruptive technology can create in a timely manner.

Although there are currently large differences between domestic and foreign markets in many areas such as large model development, application layout, business models, etc., we can see that China’s GenAI industry is constantly catching up with the international leading level, and it is expected that China’s GenAI related technologies will continue to grow in the future. and applications will gradually mature, and further explore business models suitable for their own development.

The above is the detailed content of How to make full use of McKinsey’s 10 insights into artificial intelligence?. For more information, please follow other related articles on the PHP Chinese website!