Exclusive Interview with DeepZero Intelligence CEO Huang Xiaonan
Thirty years ago, Procter & Gamble set the industry standard for marketing decision-making with its precise market research and data-backed strategies.
Today, marketing has entered the era of large models. From issuing coffee vouchers with pinpoint accuracy to real-time bidding for ads, from intelligent user segmentation to optimizing pricing strategies, the scale and variety of data involved—and the algorithms supporting them—have fundamentally changed.
Yet, the essence of marketing decision-making across economic cycles remains the same: the core business goal is still to acquire, retain, and engage customers, ultimately driving cost reduction and efficiency.
As digital marketing reaches deeper levels of complexity, businesses are grappling with a dual challenge: processing exponentially growing user behavior data while making real-time commercial decisions in an increasingly complex market.
When traditional marketing tools can no longer meet dynamic demands, AI Agent technology is opening new possibilities for “generative marketing,” enabling more precise and efficient decision-making.
As one of China’s earliest companies to apply machine learning for small-model training, DeepZero Intelligence has spent the past sixteen years pursuing a less obvious but more commercially impactful path: using AI to empower scientific decision-making for businesses.
“No other company in the B2B space understands how to integrate large models with business scenarios to optimize marketing decisions as deeply as we do,” says DeepZero Intelligence’s founder and CEO Huang Xiaonan.
Huang’s industry insights stem from her unique experience across two eras: from managing Procter & Gamble’s “small data” decision-making based on Nielsen retail audit data in the 1990s to building DeepZero’s AI-driven marketing decision system today. This gives her a deep, cycle-spanning understanding of “AI-powered decision-making.”
This CEO, who combines Procter & Gamble’s traditional marketing DNA with cutting-edge AI insights, is now leading her team to merge large language models with industry-specific small models, creating a self-evolving AI Agent decision network.
In an exclusive interview with EqualOcean, Huang Xiaonan outlined DeepZero’s latest strategic direction: integrating “large + small models” to empower marketing decisions and fully embrace AI Agents. The full interview is as follows.
From AIGC to AI Agents: AI Evolution and the Future of Marketing Applications
EqualOcean: In the past, we often heard terms like AdTech or MarTech. Recently, concepts like “generative marketing” and “AGI marketing” have gained traction. How do you view the impact of large language models’ rapid development on marketing activities over the past two years?
In the first phase of AI development, large models mainly demonstrated their capabilities through content generation, such as ChatGPT.
At DeepZero Intelligence, we focus on decision science, which is why we are more concerned with the emerging trend of AI Agents. Unlike early AIGC applications, AI Agents are more complex—they leverage large model capabilities to execute specific business workflows based on task requirements.
The value of AI Agents lies in their ability to replicate tasks previously done manually through automated processes while combining intelligent AI decision-making, significantly improving business efficiency. For example, DeepZero’s solutions go beyond providing businesses with a simple AI chatbot; we deeply integrate AI capabilities into corporate business logic, enabling a complete closed loop from insight to action.
Currently, AI Agents have become the mainstream direction for enterprise operations, and DeepZero is committed to continuous innovation in this space, providing enterprises with smarter decision-making support.
EqualOcean: What are your views on the “scenario” challenges AI marketing faces today compared to the past?
AI applications have always been scenario-driven, fundamentally aimed at solving specific pain points for specific users.
As a B2B company, we face much more complex and harder-to-define scenarios than those in B2C. For example, when ChatGPT launched, consumers quickly adopted it, but the B2B market is very different.
In B2B, you must first answer “whose problem are you solving” and then clarify “what specific problem are you solving.” Businesses have diverse needs. In DeepZero’s marketing products, we might help clients optimize user segmentation, design user journeys, or improve A/B testing decisions for their automated marketing systems.
The biggest challenge in the B2B space is ROI—whether clients are willing to pay for these features.
Thus, when adopting large models, it’s essential to meet enterprise pain points, justify spending, and ensure that returns far exceed costs. Only by satisfying all three conditions can AI achieve true scale in the B2B market.
Currently, AI’s B2B applications are still in the early stages, mostly concentrated in customer service and content generation. While these use cases are more mature, intelligent marketing decision-making is still an area under exploration.
DeepZero’s advantage lies in focusing on high-value, mission-critical decision-making scenarios, using large and small model integration to empower enterprises with more precise and efficient AI-driven marketing.
EqualOcean: DeepZero has partnered with leading companies across various industries. How do you tailor AI decision-making solutions for full-chain transformation or niche scenarios based on different client needs?
Many enterprises face their first challenge when adopting AI: clarifying their business workflows.
DeepZero has already productized corporate user operation workflows.
For example, in user operations, our products typically follow these key steps: tagging users, selecting target groups, data analysis, user journey design, and outcome evaluation. Businesses can adopt this flow to quickly implement AI and significantly enhance operational efficiency.
In terms of AI enablement, we strengthen execution effectiveness at every stage by integrating large model capabilities.
During user targeting, our AI-driven audience selection function accurately identifies potential users; in data analysis, AI provides deeper insights; and in user journey design, AI optimizes send times, content, and touchpoints for precise engagement.
For enterprises eager to embrace AI, adopting DeepZero’s full-stack user operation system allows for seamless AI-driven process implementation.
From Digital Transformation to AI Transformation: Quantifying AI-Driven Business Value
EqualOcean: What is the difference between digital transformation and AI transformation?
Previously, our industry partners focused on digital transformation. Now, we are increasingly discussing AI transformation, and we have developed a deep understanding of this process through practice.
Compared to digital transformation, AI transformation involves a deeper restructuring of a company’s knowledge systems. The core lies in integrating and applying enterprise knowledge bases.
Before deploying AI, companies must organize large amounts of internal data and knowledge, structure and segment this information, and leverage technologies like Retrieval-Augmented Generation (RAG) to enhance AI’s comprehension and decision-making capabilities, thereby truly empowering business operations.
AI transformation’s scope, the number of participants involved, and its complexity far exceed that of traditional digital transformation.
During AI transformation, enterprises will encounter new challenges, including:
Cross-department collaboration: business and technical teams must work together to refine the business logic for AI applications.
Knowledge restructuring: companies need to comprehensively organize and structure business workflows, user data, and operational knowledge.
Technical literacy: business teams must develop a foundational understanding of AI Agents to effectively collaborate with technical teams.
EqualOcean: Have you encountered questions from clients about realizing measurable value, beyond just providing know-how? How does DeepZero systematically quantify AI marketing results?
Businesses often raise questions about both “why to do it” (know-why) and “how to do it” (know-how) when implementing AI. DeepZero addresses this “soul-searching question” through concrete products and practical experience: how does AI create quantifiable business value?
We summarize our experience in the following four areas:
Increased product usage rates. DeepZero’s CDP (Customer Data Platform) and MA (Marketing Automation) products, with highly automated capabilities enhanced by AI, have significantly boosted usage rates. Previously, clients relied on traditional agencies, but now 100% of clients continuously use our products, lowering adoption barriers and expanding usage.
Cost savings. Some clients aim to improve efficiency. For example, with our AI-powered WeCom product, companies can reduce the need for customer service agents from 100 to 20, cutting labor costs significantly—this becomes a key driver for decision-making.
Higher conversion rates. We have long emphasized the integration of large and small models. In earlier small-model-driven marketing projects, clients typically saw a 30% uplift in conversion rates. After incorporating large models, we observed conversion improvements of up to 90% in some POC (proof of concept) projects.
Improved internal execution efficiency. Small-model applications previously faced adoption challenges—clients lacked familiarity, and business decisions required extra effort. DeepZero has optimized AI usability and embedding, enabling business personnel to make accurate marketing decisions through simple interactions, reducing technical barriers and boosting operational efficiency.
The “Golden Combo” of Large + Small Models: Reshaping the Marketing Value Chain
EqualOcean: How will DeepZero continue integrating large and small models to advance AI applications in marketing?
Our positioning is to empower decision-making through AI, focusing on using data and algorithms to optimize key marketing decisions.
We do not engage in creative work, media buying, traffic acquisition, or services. Instead, we specialize in leveraging AI to make precise pricing and content decisions at the moment of ad exposure, enabling AI-powered marketing.
By combining large and small models, we achieve better decision-making outcomes for clients. Currently, we are helping users leverage AI Agents to further enhance decision-making efficiency with large models.
EqualOcean: What are DeepZero’s core advantages and highlights in AI transformation for enterprises?
DeepZero’s core advantages can be summarized in three points:
First, DeepZero has deeply cultivated two core scenarios—digital advertising and user operations—for over a decade.Through small-model empowerment, we have productized these two scenarios and accumulated extensive industry experience. We have a deep understanding of critical business needs and pain points in AI applications, which forms the foundation of our strength.
Because of this, when introducing large models for decision optimization, we can achieve seamless integration and create greater value. Similar to how Salesforce successfully built Agent Force, DeepZero has a natural advantage in the marketing cloud domain, with our AlphaData platform benchmarking Salesforce.
No other B2B company understands how to integrate large models with scenarios for decision optimization as deeply as we do.
Second, even before the rise of large models, DeepZero successfully used AI technologies like machine learning to help clients improve key business metrics, gaining years of AI practice experience.
Regardless of whether it’s a small model or a large model, the core lies in effective data utilization. When applying AI, companies face challenges around data selection and model training for optimal outcomes.
In both areas, DeepZero has rich practical experience and has successfully translated AI technology into tangible business value for clients. Over the years, we have achieved significant growth in key indicators like conversion rates, giving us a unique competitive edge in the domestic market.
Third, we have deep insight into the B2B enterprise landscape.
Deploying AI Agents is not just a technical challenge but a systematic engineering project that involves multiple layers of an organization.
DeepZero’s advantage in the B2B market lies in our ability to provide comprehensive workflow solutions that help enterprises implement AI in the most efficient way.
Few companies in the market can offer such complete scenario solutions for AI Agent deployment, enabling AI to seamlessly integrate into daily business operations.
We are the largest company in the market focusing on developing the Holmes algorithm platform from the very beginning. Therefore, when providing AI solutions to enterprises, we naturally understand their needs better and can effectively help them achieve their goals.
“Out-of-the-box + Ecosystem Integration + Cost Reduction and Efficiency”: DeepZero’s Global Product Philosophy
EqualOcean: What are DeepZero’s overseas business plans and strategic adjustments for 2025?
Our overseas strategy is one of our two core strategies for 2025, alongside AI transformation.
This strategy is based on two important factors:
First, our clients are expanding overseas. Current overseas solutions vary in quality and often fail to fully meet the needs of Chinese enterprises.
When expanding abroad, Chinese companies tend to choose local suppliers like DeepZero, especially for two reasons:
Data security. User data is a core asset. For compliance and security reasons, many enterprises prefer local solutions to ensure data safety and control.
Cost advantages. Compared to overseas systems, DeepZero’s solutions are significantly more cost-effective. Our products rival international brands in functionality and performance but at a much lower total cost.
We have established branches in several markets globally, including Southeast Asia and Europe. Currently, we maintain local offices in five countries and regions, forming a complete international market support system.
Second, over the past year, our successful practices in Southeast Asia have validated our product advantages. China’s digital transformation has long been a global leader, and DeepZero’s products have been thoroughly tested and validated through long-term partnerships with leading Chinese clients.
As a result, compared to overseas brands, our products stand out in terms of design, user experience, and usability. They also offer more flexible, localized solutions.
In 2024, we deepened our presence in multiple overseas markets and successfully proved our products’ competitiveness abroad. Looking ahead to 2025, international expansion will be a key driver of our business growth.
EqualOcean: Can you provide insights for Chinese enterprises going overseas on how to leverage AI decision-making to enhance their international competitiveness?
For Chinese companies going global, AI applications abroad are often more diversified. I believe companies should focus on three key competitive advantages during international expansion:
Out-of-the-box AI capabilities. The SaaS ecosystem overseas is relatively mature, allowing businesses to quickly deploy AI Agents without complex technical setups, reducing implementation costs and accelerating go-to-market.
In China, enterprises often prefer private deployments, but overseas, cloud-based SaaS platforms enable immediate adoption.
Ecosystem integration. Companies going abroad need to deeply integrate into the international digital ecosystem. For example, DeepZero’s CDP system seamlessly connects with data from platforms like Shopify and Magento while reaching users via Facebook, Google, WhatsApp, and more. This allows businesses to efficiently link data across channels and optimize user journeys for precise marketing.
Leveraging global AI platforms. Many overseas digital platforms come with built-in AI features. Chinese enterprises can fully utilize these AI capabilities in conjunction with their marketing strategies to expand more effectively into international markets.
This differs from the more privatized ecosystem common in China.
For companies expanding overseas, the most critical factor is managing user data well and building AI capabilities based on this data—that’s the top priority.
2025: “All in AI Agent” Strategy and Global Ambition
EqualOcean: How do you view changes in AI technology this year? What breakthroughs will AI Agents achieve in the near future?
In 2024, AI applications were still in the exploration, pilot, and optimization stages. In 2025, achieving large-scale AI adoption will require further market education and accumulated experience.
What excites me most about 2025 is that AI-powered decision-making scenarios will expand significantly, continuously improving in accuracy, speed, and applicability—all built on our years of experience with data, models, and AI.
Currently, we focus on two main areas: AI-empowered marketing automation and AI-powered customer data platforms, which are the core directions for driving market growth.
We are firm believers in AI Agents and insist that AI Agents must deeply integrate with enterprises’ core products and business processes, rather than existing as standalone AI tools.
Based on this, our 2025 strategic goal is clear: “All in AI Agent”—fully investing in the development and application of AI Agents to drive the intelligent transformation of corporate digital operations.