Agents have officially entered the core processes of corporate marketing and transactions.

A membership event marketing campaign that previously required five or six roles collaborating for three to four weeks can now be prepared in just one day with the help of an Agent, saving 80% to 90% of labor. In a pharmacy's chronic disease management project, the identification of target patients increased by 100-fold, and the task execution rate of shop assistants jumped from 25% to 70%. In the member operation scenarios of chain restaurants, the sales performance activated by AI-targeted coupon distribution to potential customers was 3.1 times that of manual groups...

These efficiency improvements in companies from different industries and scales stem from the same change: AI is no longer just an assistant but has truly entered corporate marketing processes, autonomously driving the full marketing chain.

Tencent's recently released AI-Native Marketing Cloud and CloudMall 2.0 are vivid examples of this change. Unlike previous AI tools that provided isolated efficiency improvements, the AI-Native Marketing Cloud allows operators to simply set business objectives, and the system autonomously completes the entire process from insight to execution and review—Agents have officially entered the core processes of corporate marketing and transactions.

Against the backdrop of plateauing traffic and the necessity of public and private domain synergy, the AI-Native Marketing Cloud addresses three major challenges faced by AI in marketing growth: fragmented business system data, a lack of professional AI operation talent, and unclear paths to scaling (scalable) solutions due to industry customization.

01 Anxiety and Reconstruction in Marketing Growth

The management of a leading bedding brand has long been troubled by a recurring issue.

Many customers enter their stores expressing interest in purchasing a mattress priced between three to five thousand yuan, but sales associates guide them to the discount section, and the interaction ends there.

Upon analyzing extensive retail data, they discovered that many customers who mentioned a budget of three to five thousand yuan ultimately made purchases worth twenty to thirty thousand yuan. The three to five thousand yuan figure represented the "reference price for a good mattress" that consumers had seen online, not their actual spending limit.

Such misjudgments occur daily in offline stores across various brands, indicating that the experience of top salespeople is difficult to replicate among frontline operational staff. This also reflects significant shortcomings in how brands currently manage the consumer lifecycle in marketing growth scenarios.

Moreover, corporate marketing growth faces even more widespread and deep-seated anxieties than misjudging consumer demand.

On one hand, the traffic dividend is visibly reaching its peak. According to a report by the China Internet Network Information Center, by June 2025, China's online shopping user base will reach approximately 974 million, accounting for 87.9% of all internet users. With limited growth, the ROI of public domain advertising continues to decline, making it difficult to drive business growth through public domain channels alone.

At the same time, the attention spans and consumption patterns of modern users naturally span multiple platforms. However, internal marketing and growth systems within companies suffer from structural contradictions, such as fragmented data and artificially set boundaries in incentive structures. Public domain remains separate from private domain, with online and offline operations each having their own KPIs... This results in fragmented data, trapping brand operations in execution details and the need to adapt to different systems. As much as 80% of time is spent on configuring tools, aligning data, and coordinating channels, leaving little time for understanding users and designing strategies.

These realities have made the industry realize that public and private domains must be linked, and data across systems must be fully interconnected to empower frontline sales operations. This provides sufficient data and decision-making support for transaction conversion, underpinning overall corporate marketing growth.

Since the beginning of the year, the Agent craze, exemplified by OpenClaw (also known as "Longxia"), has swept the globe. In China, within just a few months, Longxia has elevated public understanding of AI to a new level—AI has transitioned from a "tool for answering questions" to an autonomous intelligent agent: perceiving the environment, orchestrating tools, executing actions, verifying results, and continuously evolving through self-improvement. This breaks through the limitations of previous digital and AI tools that provided isolated efficiency improvements while precipitate (accumulating) vast amounts of corporate experience, offering the potential to equip frontline marketing operations with an intelligent cockpit.

When market demand anxieties collide with technological capability inflection points, reconstruction becomes inevitable.

Li Xuechao, Vice President of Tencent Cloud and Head of Tencent Marketing and Transaction Products

The Tencent Marketing Cloud product development team also recognized this transformation. Li Xuechao, Vice President of Tencent Cloud and Head of Tencent Marketing and Transaction Products, told DigitInt that they began closed-door research and development before the Spring Festival, before Longxia became widely popular. They had already recognized the opportunities Agent technology brought to marketing growth.

Previously, the marketing cloud had developed various AI-powered tool applications. They observed that business processes varied significantly across industries and sectors, and applying AI on top of fixed workflows would require extensive scene customization, leading to high implementation costs.

With the advent of the Agent wave, building on market preparations and insights from early AI implementation challenges, Tencent Cloud quickly launched the Marketing Cloud MAGIC Agent 2.0 in March this year. A few days ago, at the Tencent Cloud AI Industry Application Conference, they further upgraded and released the AI-Native Marketing Cloud and CloudMall 2.0, along with an integrated marketing and transaction solution.

Unlike the past, where AI only played a supporting role in certain link (stages), the AI-Native Marketing Cloud's defining feature is its "AI-native" approach. Instead of piecing together systems around human workflows, it uses Agents to autonomously organize processes and reconstruct business chains.

02 AI-Native: Agents Reconstructing Marketing and Transactions

"For a membership event in 2025, clients configured 5-6 roles (humans using isolated AI tools) over 3-4 weeks. This year, with Agents, it was completed in just a few days." At the Tencent Cloud AI Industry Application Conference's AI Marketing and Transaction session, Zeng Wei, General Manager of Tencent Marketing Cloud Product Development, shared a real-world change experienced by a client using MAGIC Agent.

Zeng Wei, General Manager of Tencent Marketing Cloud Product Development, explaining the AI-Native Marketing Cloud product

The fundamental reason for the efficiency difference lies in the Agent technology's ability to reconstruct work methods. Last year, AI primarily provided isolated empowerment, requiring humans to operate tools and orchestrate processes step by step. Some even joked that a brand marketer's workday mostly involved "configuring systems" rather than "doing marketing." Now, Agents can autonomously orchestrate systems, with humans stepping back into the roles of goal setters and key node approvers.

This is the core distinction between Tencent Marketing Cloud's definition of "AI-Native" and the previous "+AI" model. Li Xuechao drew a clear line with a simple comparison: "The position of the '+' sign makes a huge difference; the essence is entirely different."

"+AI" means equipping humans with an assistant to make isolated efficiency improvements within existing workflows, dependent on fixed processes. "AI+", or AI-Native, means letting AI organize workflows, with humans stepping back to define goals, set boundaries, and make key decisions. It breaks through the bottlenecks of the "+AI" model—a lack of professional AI operation talent, unclear paths to scalable implementation, and fragmented business system data.

With the AI-Native Marketing Cloud, operators simply need to set business objectives, such as "activate members and promote new products as summer approaches." The AI-Native Marketing Cloud can then autonomously complete the entire marketing process, from insight (identifying consumption scenarios driven by rising temperatures) to operation (matching the most suitable activity strategies from a strategy library), decision-making (matching people, products, scenes, channels, timing, and coupons through Customer AI), execution (using CDP to segment audiences, MA to configure journeys, AIGC to generate content, and Enterprise WeChat to assign tasks), to review (feeding activity results back into the strategy library).

Operators no longer need to manually mine growth clues from vast amounts of data; AI proactively identifies opportunities, automatically matches strategies, and executes them.

Zeng Wei emphasized that the AI-Native Marketing Cloud's ability to truly integrate into corporate marketing processes relies not just on its strong execution capabilities but also on its dedicated marketing work infrastructure, Harness, which ensures it can perform tasks correctly.

Simply put, Harness does four things: enabling Agents to truly understand a company's customer data rather than analyzing it generically; ensuring Agents execute according to the company's validated operational methodologies without improvising; enabling precise matching among the six elements of "people, products, scenes, channels, timing, and coupons" instead of one-size-fits-all distributions; and ensuring stable implementation within corporate boundaries such as brand norms, budget constraints, and approval workflows. These four Harness capabilities effectively bridge the gap between large model capabilities and real corporate marketing scenarios.

Additionally, as Tencent's core platform for serving omnichannel corporate transactions, CloudMall has also undergone an AI-Native reconstruction.

CloudMall 2.0 deploys eight Agent systems, including one global hub that continuously analyzes operational data 24/7, proactively identifies issues, and insight (insights) trends, and seven vertical Agents responsible for core scenarios such as products, content, members, marketing, distribution, stores, and intelligent sales guidance. This achieves full operational process automation, from "issue identification → solution generation → execution → result feedback."

Qin Jiang, CloudMall Product Lead, explaining CloudMall 2.0's capabilities

Even more noteworthy is the synergy between products. Different stages, such as public domain acquisition, private domain operation, and transaction conversion, can be organically linked by Agents, forming a closed loop from customer acquisition to transactions.

For example, a brand might segment its target audience using CDP and push them to Tencent Advertising for public domain distribution. After users click the landing page and scan a code to join the private domain via Enterprise WeChat, the MAGIC Agent assists salespeople in personalized interactions, pushing product or activity pages from the CloudMall store. Users can then make purchases directly within the mini-program, with transaction data immediately flowing back to CDP to refine the next round of segmentation.

DigitInt learned that in industries such as footwear, apparel, and supermarkets, some companies have already achieved multi-stage Agent-driven operations—public domain acquisition, private domain operation, transaction conversion, and data feedback—based on a unified data infrastructure, further enhancing marketing transaction and business growth efficiency.

03 Growth Democratization and New Possibilities

With the implementation of AI-Native marketing and transaction capabilities, a noteworthy trend is emerging: AI's industry penetration in marketing growth scenarios is accelerating, and its path exhibits distinct characteristics compared to previous technological diffusions.

Over the past few years, marketing cloud platforms initially focused on high-ticket industries like automotive and jewelry, which had the margin space and budget for deep member operations and precision marketing. They gradually extended to medium-ticket industries like apparel and supermarkets, then to low-ticket sectors like catering. Simultaneously, industry types expanded from strong transaction-oriented product scenarios to service-centric professional scenarios like cultural tourism, hotels, and pharmaceuticals.

In the past year or two, with significant AI advancements, a counterintuitive phenomenon has emerged: some industries with lower initial digital maturity and organizational complexity have achieved rapid progress through "small yet closed-loop" application models, as technological improvements have made ROI calculations feasible. This industry scenario generalization undoubtedly represents a democratization of growth capabilities enabled by technology.

As AI in marketing and growth penetrates from high-barrier to more diverse industry scenarios, multi-dimensional changes are occurring. Efficiency improvements are just one aspect; service models and marketing philosophies are also transforming.

Jinyao Darren Tang, a 500-year-old traditional brand, noticed that with China's accelerating population aging, chronic disease management has become a major societal challenge, and traditional healthcare and pharmacy models struggle to meet the adherence and "last-mile" service demands of chronic disease management.

Since earlier this year, they have collaborated with Tencent on the "Digital Canal" project, focusing on cardiovascular and cerebrovascular chronic disease management in Darren Tang's core competency area. Using the latest AI Agent technology, they reconstructed the entire pharmacy-patient service process for combined medication scenarios in chronic cardiovascular and cerebrovascular diseases.

Previously, when a chronic disease patient entered a pharmacy to buy medication, it was typically a "one-time transaction" because human resources alone could not adequately serve each user throughout their chronic disease management cycle. Store staff also found it difficult to remember each user's medication history and management stage. Long-term health management needs, such as whether patients took their medication on time, required treatment reminders, or posed risks from combining different medications, were entirely unaddressed in traditional models.

After deploying Agents, store staff can instantly query a patient's medication history, generate medication guidance based on professional medical knowledge libraries, and push personalized health education and repurchase reminders based on "age/comorbidities/medication cycle/TCM differentiation." After incorporating Darren Tang's 500 years of accumulated knowledge on 122 exclusive product varieties and vetted medication plans into the system via Skillization, ordinary pharmacy staff gained professional judgment capabilities that were previously difficult to acquire.

"By applying AI Agents, we've upgraded traditional one-time transactions into a health service process," Ye Hui, CIO of Jinyao Darren Tang Group, told DigitInt. Meanwhile, Darren Tang's own service model is evolving toward a health management service platform.

Juewei Foods, with over 10,000 stores and 85 million members, sees structural marketing model transformations from one-size-fits-all to precision targeting in the Agent era.

Previously, they struggled to scale fine-grained operations for marketing campaigns, resorting to one-size-fits-all approaches. Coupon distributions rarely targeted specific segments, resulting in nearly identical marketing pushes to both students and families.

AI Member Agent Schematic

After collaborating with Tencent Marketing Cloud, they launched the AI Member Agent, which unifies public and private domain member assets to form over 150 tags and 1,000+ segments. The AI Member Agent integrates four Agents for audience segmentation, benefit design, intelligent product selection, and personalized content, covering the entire process from business diagnosis, audience extraction, person-product/person-coupon matching, outreach strategy design, activity execution, to performance review. Enterprises only need to "conversationally" input marketing requirements to achieve full-link automation.

This represents a shift from extensive (extensive) one-size-fits-all operations to precisely targeted intelligent marketing, updating marketing models and philosophies.

AI implementation is also driving research and development model innovations and organizational transformations. At the Tencent Cloud AI Industry Application Conference, Tang Daosheng, Tencent's Senior Executive Vice President and CEO of Cloud and Smart Industries Group, and Yao Shunyu, Tencent's Chief AI Scientist, mentioned during their dialogue that Tencent's model and product teams closely integrate and mutually enhance each other, forming a new Co-Design R&D model.

Similar interactive models are emerging as vertical AI applications land in industrial settings. Tencent Marketing Cloud's business experts and product teams embed themselves in frontline operations, linking demands with product capabilities to extract (distill) standardized, platform- precipitate (accumulatable) experiences. This accompanying approach has accelerated AI applications in vertical scenarios during Agent implementations at companies like DeRucci, Chimelong, and Chow Tai Fook.

The arrival of Agents has introduced new possibilities in marketing growth. Companies of different scales can now access equivalent marketing operation capabilities, achieving growth democratization. Within organizations, growth capabilities previously locked by experience and hierarchy are now unlocked by AI-Native products at every touchpoint and for every individual.