Today, a growing number of AI companies are seeking to deploy their products in real-world scenarios and industries. At the same time, more and more brick-and-mortar enterprises are exploring ways to empower themselves through AI.

Data from the Ministry of Industry and Information Technology shows that in 2025, China's core AI industry reached a scale of over 1.2 trillion yuan, with more than 6,200 enterprises. The National Bureau of Statistics further projects that by 2026, this figure is expected to exceed 1.7 trillion yuan.

Such rapid expansion of the industry means that a comprehensive evolution driven by mutual engagement is accelerating.

Recently, the restaurant chain Kaojiang announced its adoption of AI office solutions, deeply embedding AI into core processes such as financial invoicing, supply chain outbound auditing, and business data analysis—a vivid illustration of this trend.

Given that AI is penetrating ever more scenarios and industries, Kaojiang's decision to embrace AI comes as no surprise. What matters is understanding, through the Kaojiang example, what kind of deep empowerment traditional enterprises are initiating by embracing AI—and the corporate upgrading that follows. Only by grasping this can we truly appreciate the underlying message of Kaojiang's embrace of AI and identify the right path for deep integration of AI with the real economy.

The organizational transformation triggered by AI comes from within

Whether in the internet era or the subsequent digital technology era, we have witnessed efforts to empower and transform brick-and-mortar enterprises.

However, a closer look at these empowerment paths reveals an obvious pattern: most players approach the transformation of physical enterprises from the outside. Whether upgrading marketing tools or operating private-domain traffic, almost all remain at the level of 'external tools.' While this outside-in approach can deliver marginal efficiency gains, it has never been able to touch the core operating logic of enterprises.

After internet-based or digital transformation reached a certain stage, people found that the empowerment of physical enterprises had returned to square one.

Kaojiang's practice offers a completely different model.

Founded in Chengdu in 2013, this spicy grilled fish brand currently operates 81 directly-operated stores nationwide and has approximately 8 million members. Its flagship stores average over 9 table turnovers per day, with some staying open until the early hours. The fully direct-operated model means stores and frontline staff continue to grow, yet the number of headquarters functional department personnel cannot expand proportionally. Kaojiang CIO Luo Heng acknowledged that this has placed enormous pressure on functional departments.

It is precisely under the pressure of this structural contradiction that Kaojiang chose an inside-out path: rather than adding headcount on the front line, it installed an 'AI engine' in the back office.

Take month-end invoicing as an example. Each Kaojiang store is an independent entity, and operations personnel must apply for, download, and reconcile service fee invoices store by store across various food delivery platforms. Previously, one operations person needed over three hours to complete this. Now, operations personnel simply explain the requirements clearly; AI office connects through the browser and runs the task once, solidifying it into a reusable skill that then runs automatically on a monthly basis. Total time has dropped to 10 minutes, with only about 5 minutes of manual input.

More notably, such processes previously required finding a specialist for RPA development, but now operations personnel can build workflows themselves. This means AI not only replaces repetitive labor but also shifts the power of process design from the technology department back to the business departments themselves. The constituent elements and operating logic within the enterprise have thus undergone profound and comprehensive changes.

The changes brought by AI are profound and comprehensive

Re-examining the previous empowerment and transformation of physical enterprises through the lens of AI reveals a key limitation: enterprises often changed only externally or at the back end, without true internal renewal. Many enterprises had previously undergone digital upgrades, but those transformations were built on internet and digital technologies, and the levels they reached were ultimately limited.

When Kaojiang began embracing AI, the nature of the change underwent a qualitative shift.

In the supply chain outbound process, stores operate late into the night, and outbound operations are concentrated at night. If goods don't match outbound orders, stores may face shortages the next day. Previously, each order required 8 to 15 manual steps. After rules were written into AI skills, AI office continuously processes and periodically checks, compressing each order to 1 step and completing in about 10 minutes per day. A 'job that never ends' has become a daily automatically generated report.

Changes in business data analysis are equally significant. When business departments temporarily need data with different calibers, they previously could only submit requirements and wait for development scheduling, with delivery cycles often taking one to two weeks.

After Kaojiang connected AI office with its data middle platform, finance personnel directly submit data retrieval requests. AI assists with analysis and generates code, which is exported to Quick BI to build store dashboards. Delivery cycles have been shortened to one to two days, manual input has dropped from 50–70 hours to under 8 hours, and dashboard users have increased from over 80 to nearly 200. As of October 31, 2026, Kaojiang has accumulated over 20 reusable skills and 5 continuously running Agent workflows, available for reuse by operations, supply chain, finance, and other departments.

Around the daily operations of chain restaurants, Kaojiang is also advancing AI applications in site selection and construction, store scheduling, brand marketing, and staff training.

Ultimately, what AI brings to Kaojiang truly penetrates into the fine details within the enterprise, exerting force simultaneously from multiple dimensions.

This offers us a completely new line of thinking: AI for physical enterprises is not merely tool empowerment in a simple sense. More importantly, it begins to replace traditional enterprise elements and reconstruct a new enterprise operating logic.

The potential energy released by AI is multi-dimensional and enormous

Previous empowerment of the real economy by internet players was limited to certain aspects and certain processes, without exerting force from multiple angles, let alone bringing multi-dimensional evolution. AI's empowerment of enterprises—from front-end industries to back-end services—can find corresponding products and services to support almost every link.

Kaojiang's case precisely confirms this multi-dimensionality: invoicing belongs to financial processes, outbound auditing belongs to supply chain management, and data dashboards belong to business decision support. The three belong to different functions, different data sources, and different business logics, yet they have been reconstructed one by one under the same AI capability framework.

The 'China Restaurant AI Application Research Report 2026' jointly released by the Red Meal Industry Research Institute points out that the current AI application penetration rate in the restaurant industry is only 15%, but it is expected to rise to 50% by 2028. This means that Kaojiang's current exploration is still in a window period of extremely low penetration but steep growth. Enterprises that are first to complete AI transformation of internal processes will gain significant efficiency barriers in stock competition.

When AI begins to bring comprehensive innovation to physical enterprises across almost all processes and links, the potential energy it releases is no longer efficiency improvement in a single dimension, but a leap in the overall operating mode of enterprises.

It is precisely with this multi-dimensional and enormous release of potential energy that we have reason to believe AI can bring physical enterprises growth space far greater than that of the internet era and the digital era.

Conclusion

Kaojiang's use of AI to empower its operations can be said to be a vivid footnote to the current large-scale deployment of AI.

Unlike previous empowerment, this new empowerment of physical enterprises initiated by AI is an inside-out, profound, and comprehensive deep empowerment.

From the automatic issuance of a single invoice to the real-time verification of an outbound order, and to the instant generation of a business dashboard, changes are happening in the most everyday, most trivial, and most real links of enterprises.

For physical enterprises, riding the AI tailwind brings not just a transformation of tools, but a comprehensive renewal of the industry.