Agents have officially become integral to the core processes of enterprise marketing and transactions.

A membership festival marketing campaign, which previously required the collaboration of five or six roles over three to four weeks, can now be prepared in just one day using Agents, reducing labor costs by 80-90%. In a pharmacy's chronic disease management program, the identification of target patients increased 100-fold, and staff task execution rates soared from 25% to 70%. For chain restaurants' membership operations, AI-driven precision coupon distribution achieved potential customer activation sales performance 3.1 times higher than manual methods...

These efficiency improvements across enterprises of varying sizes and industries stem from a single transformation: AI has transitioned from a supporting role to an active participant in enterprise marketing processes, autonomously driving the full marketing chain closed loop.

Tencent's recently launched AI-Native Marketing Cloud and CloudMall 2.0 exemplify this transformation. Unlike previous AI tools that focused on isolated efficiency improvements, the AI-Native Marketing Cloud enables operators to simply set business objectives, with the system autonomously managing the entire process from insight to execution and review. This marks the official entry of Agents into the core of enterprise marketing and transactions.

Amid stagnating traffic growth and the inevitability of public-private domain integration, the AI-Native Marketing Cloud addresses three major challenges that AI previously faced in marketing growth: fragmented business system data, a shortage of professional AI operations talent, and unclear paths to scalability due to industry-specific customization.

01 Anxiety and Reconstruction in Marketing Growth

The management team of a leading bedding brand has long faced a dilemma.

Numerous customers enter their stores expressing interest in purchasing mattresses priced between 3,000-5,000 RMB, yet sales staff often guide them to discounted sections, where the process abruptly ends.

Upon analyzing extensive retail data, they discovered that many customers who mentioned a 3,000-5,000 RMB budget ultimately made purchases worth 20,000-30,000 RMB. The 3,000-5,000 RMB figure represented a 'reference price for good mattresses' that consumers had seen online, not their actual spending limit.

Such misjudgments are common in offline stores across various brands, highlighting the difficulty of replicating top salespeople's expertise across all frontline staff and revealing significant shortcomings in current brand-consumer lifecycle management within enterprise marketing growth.

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

On one hand, traffic dividends are visibly reaching their 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 potential, the ROI of public domain advertising continues to decline, making business growth difficult to sustain through public domain channels alone.

At the same time, while ordinary users' attention spans and consumption patterns naturally span multiple platforms, enterprises' internal marketing and growth systems suffer from structural contradictions such as fragmented data and artificially imposed boundaries in incentive structures. Public domains remain separate from private domains, with online operations following one set of KPIs and offline operations another... This results in data silos, trapping brand operations in execution details and system adaptation needs, with 80% of time spent on tool configuration, data alignment, and channel coordination, leaving little for user insight and strategy design.

These realities have made the industry recognize the necessity of integrating public and private domains while fully connecting inter-system data to empower frontline sales operations with sufficient data and decision support, thereby sustaining overall enterprise 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 new heights—shifting AI's role from a 'tool that answers questions' to an autonomous intelligent agent capable of sensing environments, deploying tools, executing actions, verifying results, and continuously self-evolving during operation. This breakthrough transcends the limitations of previous digital and AI tools focused on isolated efficiency gains while preserving vast amounts of enterprise 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 DigitFront that they began closed-door research and development before the Spring Festival, before Longxia gained widespread popularity, as they had already perceived 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, making customization for fixed workflow-based AI applications costly to implement.

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

Unlike past AI applications that only played supporting roles in specific areas, the AI-Native Marketing Cloud's defining feature is its 'AI-native' design. Rather than piecing together systems around human operations, it reconstructs business processes through autonomous Agent organization.

02 AI-Native: Agents Reconstruct Marketing and Transactions

'For a membership festival event in 2025, clients needed 5-6 roles (humans using isolated AI tools) working for 3-4 weeks. Recently, using Agents, the same task was completed in just a few days,' shared Zeng Wei, General Manager of Tencent Marketing Cloud Product Development, at the Tencent Cloud AI Industry Application Conference's AI Marketing and Transactions session, highlighting a real transformation achieved 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 this efficiency difference lies in the Agent technology-driven reconstruction of work methods. Last year, AI primarily provided isolated empowerment, requiring humans to operate tools and sequence processes step by step for configuration. Some even joked that brand marketers spent most of their days 'configuring systems' rather than 'doing marketing.' Now, Agents autonomously connect systems, with humans stepping back into roles as goal setters and critical approval points.

This distinction defines the core difference between Tencent Marketing Cloud's 'AI-Native' approach and previous '+AI' models. Li Xuechao succinctly clarifies the boundary: 'The position of the '+' sign makes a huge difference—the essence is completely different.'

'+AI' means equipping humans with an assistant for localized efficiency improvements within existing workflows, dependent on fixed processes. 'AI+', or AI-Native, lets AI organize workflows while humans define goals, set boundaries, and make critical decisions. This breakthrough addresses precisely the bottlenecks of the '+AI' model: lack of professional AI operations talent, unclear paths to scalable implementation, and fragmented business system data.

With the AI-Native Marketing Cloud, operators simply set business objectives, such as 'activate members and promote new products as summer approaches.' The system then autonomously handles the full marketing process: insight (identifying consumption scenarios triggered by rising temperatures), operations (matching optimal activity strategies from a strategy library), decision-making (using Customer AI to align people, products, venues, channels, timing, and coupons), execution (deploying CDP for audience segmentation, MA for journey mapping, AIGC for content generation, and Enterprise WeChat for task assignment), and review (feeding activity results back into the strategy library).

Operators no longer need to manually mine growth clues from massive datasets—AI proactively identifies opportunities, automatically matches strategies, and executes them.

Zeng Wei emphasizes that the AI-Native Marketing Cloud's ability to truly integrate into enterprise marketing processes relies not just on its robust execution capabilities but also on its dedicated marketing work foundation, Harness, which enables Agents to perform effectively and accurately.

Simply put, Harness accomplishes four key tasks: enabling Agents to truly understand enterprise customer data rather than conducting superficial analysis; ensuring Agents execute according to enterprises' verified operational methodologies without improvisation; enabling precise matching among the six elements of 'people, products, venues, channels, timing, and coupons' instead of one-size-fits-all distribution; and ensuring stable implementation within enterprise boundaries such as brand guidelines, budget constraints, and approval workflows. These four Harness capabilities also effectively bridge the gap between large model capabilities and real enterprise marketing scenarios.

Additionally, as Tencent's core platform for enterprise omnichannel 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 detects trends, along with seven vertical Agents responsible for core scenarios such as products, content, membership, marketing, distribution, stores, and intelligent sales guidance. This enables full operational process automation from 'issue identification → solution generation → execution → result feedback.'

Qin Jiang, CloudMall Product Lead, explaining CloudMall 2.0 capabilities

Even more noteworthy is the synergy between products. Agents can organically connect different stages such as public domain acquisition, private domain operations, and transactional conversions, forming a closed loop from acquisition marketing to transactions.

For example, a brand might use CDP to segment target audiences and push them to Tencent Advertising for public domain distribution. After users click landing pages, scan codes, and join Enterprise WeChat for private domain engagement, MAGIC Agent assists sales staff in personalized interactions, pushing CloudMall product or activity pages. Users then make direct purchases within mini-programs, with transaction data immediately flowing back to CDP for more precise next-round segmentation.

DigitFront has learned that in industries such as footwear, apparel, and supermarkets, some enterprises have already achieved multi-stage Agent-driven operations—public domain acquisition, private domain operations, transactional conversions, and data feedback—based on a unified data foundation, 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 notable trend is emerging: AI's industry penetration in marketing growth scenarios is accelerating, following a path distinct from previous technological diffusion patterns.

Over the past few years, marketing cloud platforms initially focused on high-ticket industries like automotive and jewelry, which had both the incentive and budget for deep membership operations and precision marketing. Services gradually extended to medium-ticket sectors like apparel and supermarkets, then to low-ticket areas like catering. Simultaneously, industry coverage expanded from strongly transactional product scenarios to service-centric professional fields like cultural tourism, hotels, and pharmaceuticals.

In the past year or two, however, 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 advancements 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 gains are just one aspect—service models and marketing philosophies are also transforming.

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

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

Previously, a chronic disease patient's pharmacy visit represented a typical 'one-time transaction,' as human resources alone couldn't adequately serve each user throughout their chronic disease management cycle, nor could staff easily track each user's medication history and management stage. Long-term health management needs such as medication adherence, treatment reminders, and risks of combined medication use remained unaddressed under traditional models.

Once the Agent is deployed, pharmacy staff can promptly access patients' medication histories, create medication guidance based on professional medical knowledge bases, and send out personalized health education materials along with repurchase reminders. These reminders are tailored according to factors such as 'age, comorbidities, medication cycle, and Traditional Chinese Medicine (TCM) differentiation.' By integrating Daren Tang's 500 years of accumulated expertise on 122 exclusive products and verified medication plans into the system as Skills, ordinary pharmacy staff now possess professional judgment capabilities that were previously challenging to acquire.

'By leveraging AI Agents, we've transformed traditional one-time transactions into a comprehensive health service process,' Ye Hui, CIO of Jinyao Daren Tang Group, informed DigitFront. Concurrently, Daren Tang's service model is evolving into a health management service platform.

Juewei Foods, with over 10,000 stores and 85 million members, recognizes the potential for structural transformation in marketing models within the Agent era—transitioning from a one-size-fits-all approach to precision targeting.

Previously, their marketing campaigns faced challenges in implementing scalable refined operations, often resorting to undifferentiated audience segmentation. Coupon distribution rarely targeted specific groups, leading to nearly identical marketing messages being sent to both students and families.

AI Member Agent Schematic

After collaborating with Tencent Marketing Cloud, Juewei Foods launched an AI Member Agent that unifies public and private domain member assets, generating over 150 tags and more than 1,000 audience segments. The system 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 simply need to input marketing requirements through a 'conversational' interface for full-chain automated execution.

This marks a shift from crude, undifferentiated operations to precisely targeted intelligent marketing, with marketing models and philosophies undergoing a renewal.

The implementation of AI is also driving innovation in R&D models and organizational transformation. During the Tencent Cloud AI Industry Application Conference, Tang Daosheng, Tencent's Senior Executive Vice President and CEO of Cloud and Smart Industries Group, discussed with Yao Shunyu, Tencent's Chief AI Scientist, how Tencent's integration of models and products has fostered a new Co-Design R&D approach.

Similar interactive models are emerging as vertical AI applications penetrate industrial sites. Tencent Marketing Cloud's business experts and product teams work on the front lines, connecting requirements with product capabilities to extract standardizable, platform-accumulated experience. This accompanying approach has accelerated AI applications in vertical scenarios during Agent implementations at companies like DeRucci, Chimelong, and Chow Tai Fook.

The advent of Agents has opened up new possibilities for marketing growth. Enterprises of all sizes now have access to equivalent marketing operational capabilities, achieving growth democratization. Within organizations, growth capabilities that were previously restricted by experience and hierarchy have now been unlocked through AI-Native products, empowering every touchpoint and individual.