AI Trends

One Person Can Deliver, Three Can Match a 20-Person Team: “AI + Advertising” OPCs Proliferate in Shanghai’s Xuhui District

Shanghai ObserverUpdated Nov 19 09:33

AI summary

“AI + advertising” OPCs are proliferating in Shanghai’s Xuhui District, which has now attracted and incubated more than 100 advertising-related OPCs. A three-to-five-person team can generate as much video content as a 15-to-20-person team at a traditional advertising company. The district has established the city’s first alliance for taking AI advertising global, as well as an “AI + Advertising Industry Comprehensive Service Center,” providing businesses with services including registration, preliminary compliance reviews, resource matchmaking, and project incubation.

Body · AI translation

At the recently held 2026 AI Advertising Innovation Season and “AI + Advertising” Business Partner Program showcase, the city’s first consortium for taking AI advertising overseas was officially established. In Xuhui District, where the event was held, advertising industry revenue exceeded RMB 56.8 billion in the first 10 months of this year, maintaining double-digit growth. The district’s eight interconnected Super Entrepreneur Communities have collectively attracted and nurtured more than 100 advertising-related OPCs (one-person creative companies).

As “AI + advertising” increasingly shows the potential to rival the traditional advertising industry, a widely discussed question is emerging: Now that AI has lowered the barrier to video generation to the point where anyone can try it, how can the AI video generation industry avoid homogeneous competition, and how can it ensure that technology genuinely serves content innovation instead of fueling repetitive, cutthroat competition?

Three people doing the work formerly done by 20

In the West Tower of West Bund Smart Tower, where Shi Jiaqi, head of SenseTime’s Seko industrial base, works, OPC teams move in to work every day. He described the smallest viable startup unit: a solo entrepreneur goes out to negotiate a partnership and, once an agreement is reached, uses AI software to manually craft and deliver the visual product. But this is not the most advanced form of entrepreneurship found in OPC communities.

“Top-tier OPCs possess creative capabilities themselves. When a client presents a request, the entrepreneur can develop the concept and write the script independently while also handling color, voice-over, music and screenwriting.” Such OPC teams typically consist of three to five people, yet their video-generation capabilities are equivalent to those of a 15- to 20-person video team at a traditional advertising agency.

One knock-on effect of this leap in efficiency has been a surge in entrants to the “AI + advertising” sector. “The barrier to getting started with AI tools is relatively low, so more and more people are entering the industry,” Shi said candidly. As with micro-short dramas, once the industry’s barriers fell, quoted prices continued to decline and various forms of disorderly competition emerged. Yet amid the apparent free-for-all in “AI + advertising,” the industry has also seen another side: advertising formats and distribution platforms are far more diverse than before. Ads now range from app launch screens to product placement in dramas, and appear on platforms from Douyin to Xiaohongshu; sometimes viewers do not even realize they have watched an advertisement.

The “AI Advertising Festival” jointly organized by SenseTime and Tuhu Car Maintenance is one example. With no restrictions on scripts or subject matter, creators can enter by producing story-driven videos, posting them on social platforms and interacting with the brand’s official account, with prizes including cash and vehicle maintenance services. “AI advertising does not require heavy-handed product placement in the content itself. In a love story, for example, the male and female leads could meet for the first time while having a car repaired at Tuhu. All of this can be generated with AI.”

Xinpianchang, a video-creator platform with 15 million registered users, established a new market entity in Xuhui this year. Its head, Chen Yue, divides current advertising projects into two categories: high-budget projects involving celebrity shoots or demanding a premium visual texture are still primarily filmed in the real world, with AI mainly used to assist with 3D special effects and reduce costs; lower-budget advertisements with less stringent requirements for fine detail are already more inclined to use entirely AI-based production.

“Competition in the industry is currently intense, and both the technology and the industry are in a transitional period. Teams with signature AI works are better positioned to capture market recognition, but ultimately the core competitiveness of an advertisement still depends on its production standards and quality,” Chen said.

Human aesthetic judgment may fall below AI’s

As industry barriers fall and supply surges, what standard can distinguish the good from the mediocre? In interviews, industry insiders all pointed to the same factor: people.

Shi said bluntly that creativity remains the core of the “AI + advertising” industry, followed by aesthetic judgment and the ability to make decisions about AI. He noted that AI can already meet the passing grade for aesthetics by human standards, while some people’s aesthetic judgment may actually fall below that of AI. If creators’ abilities extend no further than knowing how to use AI tools, the content they produce may not even reach the average standard of content generated by AI in “fully automatic” mode.

One of SenseTime’s responses has been to establish a strategic partnership and an innovation and entrepreneurship base with the Shanghai Theatre Academy. It offers credit-bearing AI courses that expose students in traditionally hands-on disciplines such as directing and screenwriting to cutting-edge AI tools, and then identifies entrepreneurs with professional foundations who can join OPC communities.

Chen said that Xinpianchang draws on its platform community and its “Shotlab” AI canvas tool to match creators with brands more precisely while helping them achieve commercial deployment.

There are three specific methods of matching supply with demand: brands can publicly search creators’ work; clients can enter the type of advertisement and their budget, after which the system recommends a team and generates a shot-by-shot storyboard; or they can use official production services. In Chen’s view, AI can provide basic output, while strong aesthetic judgment can guide it to generate higher-quality content, thereby making “AI + advertising” an industry that leads technological innovation and content creativity rather than becoming a bubble.

SenseTime has gone a step further. Shi said the company signs outstanding creators into the SenseCore ecosystem as “contracted directors” and has also invested in teams capable of producing both short dramas and advertisements. This series of initiatives stems from the fact that entrepreneurs in AI video often lack traditional “investment credentials,” such as elite university backgrounds, patents and academic papers. Consequently, while investors’ understanding of the emerging “AI + advertising” field remains incomplete, platforms and leading industry companies are filling the gap in talent development.

More “enjoy without applying” policies and less legwork for businesses through data sharing are needed

Cultivating an industrial ecosystem requires the government to understand the industry deeply. A relevant official from the Xuhui District Administration for Market Regulation said Xuhui is shifting from being a traditional “administrator” to a “development partner” for enterprises, accelerating the formation of a new industrial landscape in which “the government provides the platform, industry leaders take the lead and the entire value chain collaborates.”

The city’s first “AI + Advertising Industry Integrated Service Center,” inaugurated in Xuhui that day, brings together functions including business registration, preliminary compliance reviews, data intellectual property registration, resource matching and project incubation. An inaugural group of 40 members was simultaneously assembled into an AI advertising mentor team to provide young creators with professional guidance and hands-on coaching.

However, bottlenecks remain in policy implementation. Shi has observed three problems facing AI content entrepreneurs: micro and small enterprises lack dedicated staff to handle complicated application materials; companies must still proactively monitor various policy announcement channels and can easily miss policy updates; and application templates differ across government departments, whose data systems are not fully interconnected, resulting in repeated form-filling. He suggested expanding “enjoy without applying” policies for OPCs, using big data to automatically identify eligible companies and immediately notify them of new policies, moving the entire process online and enabling data sharing among departments so that businesses can truly “let the data do more of the legwork.”

Chen noted that platform creators, like the platform itself, are concerned about computing costs and ecosystem connections. Xinpianchang is paying close attention to support policies covering Token consumption and computing-power vouchers in order to reduce the operating costs of AI platforms. The platform also hopes the government will provide more spaces for industry practitioners to connect and introduce entrepreneurs to high-quality clients.

The reporter learned that as of October, Xuhui had held 21 related industry events this year, establishing deep connections with more than 120 brands, over 400 advertising companies, more than 50 large-model companies and over 400 university creators. Xuhui is supporting every stage of entrepreneurship with a systematic service network and bringing together creative resources from around the world through an open ecosystem.

Source

Shanghai Observer