Why Does a Product Manager Have to Be Human? LynxAI Breaks the New Product Curse with Three Agents in Cross-Border E-commerce | AI Competitiveness Interview

【Ebrun Original】In the past, cross-border e-commerce companies often faced problems such as low success rate of new products, long product development cycles, high talent dependence, and insufficient market insight when planning products. However, now, in the current fast commercial application of AI technology, these are no longer issues. Focusing on AI innovation service provider in the cross-border e-commerce field, LynxAI has achieved a systematic solution for cross-border e-commerce sellers by providing market opportunity insights, product innovation combination optimization, and systematic market operation strategies through its self-developed product innovation model and Agentic AI framework, based on the "AI Product Manager."

Their "AI Product Manager" application platform includes several professional Agents, such as "Market Insight Expert Jaxx," "Product Planning Engine Jobs," and "Commercialization Strategist Pony." Among which, "Market Insight Expert Jaxx" scans the industry value chain dynamically and discovers new product opportunities on a real-time basis through insights into consumers, the market, the supply chain, and technology; "Product Planning Engine Jobs" conducts comprehensive inference from multi-source information to carry out combined product innovation and definition, thereby shortening the product development cycle and increasing the product success rate; and "Commercialization Strategist Pony" is based on the four-dimensional insights to establish a clear product commercial success plan, enhancing the precision and efficiency of product sales and operations.

With the help of these Agents, the operational methods of cross-border enterprises have been restructured: first, it helps companies transition from the previous trend-following mode of finding the same style on 1688 based on the platform's explosive data to a first-innovation mode based on deep insight; second, it integrates consumer data and trends from multiple global markets, helping companies establish a global perspective across platforms and markets; in addition, it transforms the traditional decision-making process that relies on personal experience into scientific decision-making based on multi-dimensional data analysis, making the product manager's implicit knowledge explicit and building a shared knowledge base and innovation methodology for the enterprise.

Focusing on the theme of "What kind of AI competitiveness do global new brands need," Ebrun invited many cross-border service companies to conduct a series of interviews to share how they apply AI technology to enhance the level of service for overseas enterprises and help them improve their overall competitiveness. The following is the interview content of Ebrun with LynxAI's co-founder and product manager, Ding Mingming, after editing.

Ebrun: Please introduce your company and main business.

Ding Mingming: LynxAI is a startup company specializing in the application of AI. We are dedicated to using artificial intelligence technology to solve the problem of product innovation in the cross-border e-commerce field. The company name "LynxAI" is derived from the lynx, symbolizing that although we are a small entrepreneurial team, we possess the agility and hunting ability like a lynx.

Our main business is to provide cross-border e-commerce enterprises with AI-based product innovation solutions, including systematic services such as market opportunity insights, product innovation combination optimization, and market operation strategies. Through our self-developed product innovation model and Agentic AI framework, we have built an AI intelligent body that can think and work like a product manager, helping companies reduce the risk of new product development and improve product innovation efficiency.

Our founding team consists of professionals with backgrounds in cross-border e-commerce, product management, and AI technology:

Founder Ke Yi previously served as COO of Zhio Home, Chief Strategy Officer of Anker Innovation, and General Manager of Strategy Execution at Fosun Group;

Co-founder Ding Mingming is responsible for product success, and has previously served as a digital business partner at Oookini, co-founder of Guoke, and head of mobile business at Tudou.com;

Chief AI Scientist Liu Tong has a cross-disciplinary background in applied mathematics, marketing, computational advertising, and human-computer interaction;

CTO Li Xiaoteng previously served as a high-performance technology expert for Baidu Mobile Search, head of product research and development at Blued, and Chief Technology Officer of AIUX.

Ebrun: What are your products and their core application scenarios? How do they help global brands/enterprises improve their competitiveness in going global?

Ding Mingming: Our core product is a set of "AI Product Manager" application platform, which includes multiple professional Agents:

1. Global Market Insight Expert "Jaxx": This is our first product that we will officially release in May 2025, mainly used for market opportunity analysis and consumer insights.

2. AI Product Manager "Jobs": Responsible for product planning and definition, helping companies develop and optimize new products.

3. AI Product GTM "Pony": Responsible for product commercialization strategy, helping companies formulate market deployment and operation plans.

These Agents are mainly used by cross-border enterprises for annual product strategic planning, new product development, product combination optimization, market opportunity response, and other scenarios, helping cross-border enterprises solve the following key pain points:

1. Low success rate of new products: According to TikTok data, the proportion of explosive products in the overseas market is as low as less than 0.6%. Through AI-driven market insights and consumer understanding, we help companies improve the success rate of new products.

2. Long product development cycles: Traditional market research and consumer insights on average take 2 weeks. Our products can increase the speed of response to the minute level, helping product managers complete the same analysis work within 2 hours.

3. High talent dependence: Excellent product managers cultivated by companies for many years may leave at any time. Our AI product manager can precipitate the knowledge base of the company and reduce reliance on individual capabilities.

4. Insufficient market insight: It is difficult to collect social media and consumer data in overseas markets, and the traditional third-party consulting costs are expensive (ranging from tens of thousands to millions of RMB). We provide more economical and efficient intelligent insight solutions.

We help cross-border enterprises restructure their operational methods:

1. Transitioning from imitation to innovation: We help companies shift from the previous trend-following model of finding the same style on 1688 based on platform explosive data to an innovation-first model based on deep insight.

2. Transitioning from a single platform to a global perspective: Integrating consumer data and trends from multiple global markets, helping companies establish a global perspective across platforms and markets.

3. Transitioning from experiential decision-making to data-driven: Transforming the traditional decision-making process that relies on personal experience into scientific decision-making based on multifaceted data analysis.

4. Transitioning from isolated work to knowledge sharing: Making the implicit knowledge of product managers explicit, building an enterprise's shared knowledge base and innovation methodology.

Ebrun: How has the rapid development of large-scale model and generative AI technology brought changes/upgrades to your products?

Ding Mingming: The rapid development of large-scale model and generative AI technology has brought revolutionary changes to our products:

First is the reshaping of the development method:

1. Small team, great capability: Today, AI technology enables small teams like ours to develop powerful AI applications. This completely changes the traditional model that requires thousands of R&D personnel in IT companies.

2. Modelization of complex knowledge: We can transform the implicit knowledge and best practices accumulated by product managers over the years through the MOE hybrid expert model and Agentic AI framework into AI workflows that can handle complex tasks.

Second is the upgrade in product capabilities:

1. Multi-dimensional cross-inference: Based on the inferential capability of large-scale models, we have achieved four-dimensional cross-analysis of market insight, consumer insight, supply chain insight, and technology insight, which is difficult to achieve with traditional analysis tools.

2. Nonlinear correlation mining: Through the generative AI capability, we can discover hidden correlations between data, such as the explosion of the hot topic "the first batch of cats returning to the village for the Chinese New Year," which may reflect the outbreak of pet travel scenarios and changes in human-pet relationships.

3. Autonomous continuous learning: Our AI intelligent body can undergo self-learning reinforcement based on data and user feedback, continuously improving delivery quality, which is the advantage of Agentic AI compared to traditional data tools.

In addition, there is a change in the interaction method. It changes from traditional form filling and report query to natural language conversation. Product managers only need to input a set of Asin List (Amazon product ID), and Jaxx can carry out comprehensive market analysis.

In the future, we plan to collaborate with large AI application platforms, embed them into enterprise IM systems, and make the AI intelligent body more like a real AI colleague, ready to respond to team members' inquiries and tasks.

Ebrun: How did your technological and product advantages accumulate?

Ding Mingming: LynxAI has a team with a cross-disciplinary background—experienced managers in the e-commerce industry and AI experts. This cross-disciplinary composite team background makes customers more willing to open their operational data and allows us to deeply combine technology and business. Meanwhile, our Chief AI Scientist, Liu Tong, has a cross-disciplinary background in applied mathematics, marketing, computational advertising, and human-computer interaction, enabling the design of intelligent rules and algorithms that better meet business needs.

At the organizational level, we have separated content engineering and Agent teams from the architecture of AI technology engineering, led by professionals with backgrounds in consumer market research, computational advertising, and consumer behavior research. Through professional content engineering, we ensure the quality of AI output and enable it to express itself in a real product manager's tone, rather than the programmer-style writing commonly seen in large models.

At the same time, we have established a deep cooperative relationship with several top cross-border e-commerce companies, continuously verifying and optimizing products through real business scenarios.In addition, there have been innovations in technical architecture, including: the self-developed MOE hybrid expert model—designed to meet the special requirements of product innovation, this model combines domain models/algorithms and business rule libraries; multi-level intelligent architecture—built from the underlying LLM (Cloud/Qwen, ChatGPT, Deepseek, etc.) to professional reasoning capabilities, and then to the expert agent's multi-level intelligent architecture; intelligent data processing flow—developed a complete process for intelligent acquisition of consumer data, industry rule processing and cleaning, and multi-round recursive reasoning.

Ebrun: From the perspective of the industry segment you are in, how effective and extensive is the application of AI technology and tools for global brands/overseas enterprises? What changes and challenges are involved in this process?

Ding Mingming: As pointed out by Yang Binglong, founder of HJF Group, "Cross-border brands are generally accustomed to driving efficiency improvement with technology, thus gaining greater competitiveness. Therefore, they are more open and even eager towards AI technology from the beginning."

Currently, the application scope is expanding from marketing to product development—initially AI was mainly used in marketing scenarios in the cross-border e-commerce field, but now more and more enterprises hope that AI can help discover global market opportunities and assist in product definition. Moreover, compared to traditional SaaS tools, AI tools have a simpler interaction and lower customer training and education costs.

However, in the process of applying AI, cross-border enterprises also face some challenges. For example, business model transformation—how to transform from the previous model of relying on "store groups + stocking" to product polishing, new product development, and brand incubation; knowledge accumulation and inheritance—how to transform the successful experience in the product manager's mind into enterprise knowledge assets, to reduce the impact of talent flow; data quality and processing—processing and screening of massive unstructured data (such as comments, social media content), extracting valuable insights rather than surface information; AI output quality—how to ensure the professionalism and practicality of AI output content, and avoid vague suggestions and impractical solutions; balance of cost and return on investment—balancing the investment in AI application with the actual business value, especially for small and medium-sized enterprises.

Ebrun: What are your observations on the application of AI in the field of brand globalization/cross-border e-commerce? What trends do you foresee in the coming years?

Ding Mingming: Firstly, AI technology is bringing cognitive equality to enterprises, enabling more small and medium-sized enterprises to obtain market insights and product innovation capabilities that were previously only affordable for large enterprises. The cognitive equality brought by AI can break through the innovation bottleneck and reshape the business world by breaking the value chain of innovation.

Secondly, the application of AI in the field of cross-border e-commerce is shifting from simple efficiency tools (such as translation, customer service) to deep decision support (product innovation, market strategy).

In addition, general large models are insufficient to meet specific industry requirements, and specialized AI intelligences for specific scenarios are emerging in various industries.

As for the forecast future trends:

1. Shift of competitive focus: as Yang Binglong predicted, "With the widespread application of AI capabilities in the field of cross-border e-commerce, the competition between enterprises in the future may shift from product R&D and business opportunity exploration to the competition in decision-making power, brand power, and organizational efficiency."

2. Multi-agent collaboration ecosystem: the development of multi-agent collaborative systems from single AI assistant to different AI intelligent agents collaborating to complete complex tasks, for example, our product system includes AI insight expert Jaxx, AI product manager Jobs, and AI product GTM Pony and other expert agents collaborating.

3. Upgraded human-machine collaboration mode: the future will form a more efficient human-machine collaboration mode, with AI responsible for data analysis and solution generation, and humans responsible for creative guidance and final decision-making. Zhang Keyi believes that "in the future, people may only need to do two things—invent and consume, that is, innovate and verify innovation, and all the work in the middle of this chain will be achieved by AI intelligent agents."

4. Deepening integration of industry knowledge: the application of AI will more deeply integrate industry expertise, data, and technological capabilities, forming a truly "expert-level" AI system.

5. Platformization and standardization: the application of AI will move towards platformization and standardization, lowering the usage threshold and enhancing configurability, enabling small and medium-sized enterprises to easily apply AI.

6. Enhanced localization capability of global brands: AI will help global brands better understand the cultural differences and consumer habits in regional markets, achieving more precise localization strategies.

Our company's slogan is "No need for 1000 eyes to tame complexity", and we believe that through the application of AI technology, more cross-border e-commerce enterprises can solve complex operational problems and achieve the democratization of innovation, thereby gaining greater competitive advantages in the global market.

AI is entering commercial reality, and brands and cross-border e-commerce are taking the lead. The 2025 Global New Brand AI Competitive Conference is scheduled for June 6th in Hangzhou with the theme "Evolutionary Choices", jointly organized by Ebrun and JiaYu Capital. Stay tuned for updates: