Survival Guide for Product Managers in the AI Wave
-
With the rise of the concept of AI product managers, people often ask: What’s the difference between an AI product manager and a regular product manager? Aren’t they both doing demand research and product design? Why does AI make it so different?
This is the question I want to explore today: How has AI transformed the role of product managers?
We’ll discuss this from four perspectives:
In recent years, several events have brought artificial intelligence into the public spotlight.
After two years of development, AI has left its mark everywhere in our lives. Like the advent of electricity a century ago, AI has revolutionized every product, seemingly reshaping the way everyone lives.
There have always been two types of negative voices:
In reality, we use AI in our daily lives without even realizing it. John McCarthy, who first coined the term 'Artificial Intelligence' in 1956, often complained, 'As soon as something works with AI, people no longer call it AI.'
Because of this effect, AI always sounds like a mysterious future concept rather than an existing reality. At the same time, it makes people think AI is an unrealized trend.
Personally, I believe AI impacts the internet in two ways: 'enhancing user experience' and 'improving decision-making efficiency.'
Here’s a common example:
As Yu Jun once said, changes in the product management industry are always accompanied by the integration of new elements into human life. Although AI emerged last century, it is now transforming product managers in a whole new way.
In the traditional internet era, due to the红利 of traffic effects, product managers focused mainly on挖掘 user needs and improving user experience.
With the development of AI, a细分领域 has emerged in the product management industry: AI product managers. Their work revolves around applying AI technology to scenarios, using AI as a powerful tool to solve problems and enhance product functionality for better user experience and efficiency.
While both roles本质上 aim to meet user needs, there are significant differences in思维方式 and解决方式, as seen in the following two aspects:
In the internet era,流量为王, and 'user-first' became the consensus among互联网企业. Whether in e-commerce or social media, product managers studied user behavior daily to挖掘 user needs and create satisfying products.
In the AI era,数据为王, and data has become the compass for businesses.
Our products and users can be described with data,而不是 relying on经验主义 to guess what users want. Now, we have more concrete, quantifiable ways to判断 customer needs.
The diagram below explains the hierarchy and importance of data, information, and knowledge. All decision-making knowledge must be基于 information. Decisions made solely on intuition, without data支撑, cannot积累沉淀 into knowledge. Some companies collect data but don’t know how or where to use it.
Data sitting idle has no value. Effective data-driven strategies can fully transform企业 data into information and结构化知识体系, efficiently guiding rapid business development.
As Dr. Deng Xiong said: AI represents a major变革. We no longer see users as the center围绕 which一切工作 revolve. Instead, we turn users into data, transforming all user behavior into data, and reflecting user performance in products through data. This model is the result of数据思维导向.
Traditionally, there were two ways to innovate in products. One was to借鉴 knowledge from another field, like how婴儿恒温箱 was inspired by动物园恒温箱. The other was to引入跨行业的新要素, turning a product into a new species, like朵亚朵亚酒店 using众筹 to create loyal customers.
Today, AI technology upgrades have also elevated product managers'认知升级. The变革 in product solutions is颠覆式创新—not just重组旧元素 but upgrading production factors with AI.
For example, the essence of the运输行业 is solving时效问题. The shift from马车 to汽车 improved speed by enhancing动力. Today,动力 isn’t the main限制因素; safety and driver fatigue are. Autonomous trucks use AI to reduce人为因素, making 'human drivers' less critical.
This认知升级 shows how product managers rethink scenarios from new dimensions,不再 relying on旧因素 for efficiency but upgrading them in new ways.
Most AI products today are落地 applications of decades of machine learning and深度学习 research. While技术上 has突破, algorithms still have limitations, excelling only in单点优化.
Thus, AI products are characterized by精细化,个性化,智能化, and多模态. They are highly intelligent and人性化 in specific fields and scenarios.
The快递行业, with its标准化流程, high人力成本, and low分拣效率, is a key场景 for AI. Traditionally, designing a通用型分拣装置 for all快递 types is nearly impossible due to high system demands.
But can we make no improvements? Not at all. Focusing on细分场景 like小件普通物品, AI can significantly提升物流效率. Trying to满足通用需求 upfront may not help. Only by细分不确定性 can we match technology with场景 and evolve iteratively.精细化建设 is the hallmark of AI products.
Traditionally, human-computer interaction followed a rigid command-response pattern. But in the AI era, we can create more imaginative scenarios through divergent, personalized interactions.
When searching for movies, users primarily want to find films they'd enjoy rather than navigating complex filters. Previously, we could only narrow down by genre before manually browsing synopses and reviews—an inefficient process, especially when users lack clear preferences. This remains a core challenge for product managers.
AI enables natural queries like "mystery movies with plot twists" or "the wolf-hunt scene in The Longest Day in Chang'an." It can even recommend films based on users' current moods—mirroring conversations with movie-savvy friends who anticipate preferences. This personalization distinguishes AI products from traditional ones.
Financial institutions once struggled with manually auditing call center compliance through random sampling—a reactive, inefficient approach. Now, AI-powered speech recognition analyzes all calls in real-time, preventing violations proactively. This exemplifies AI's intelligent value.
The concept of "modality," originally from biology, refers to sensory channels like vision or hearing. Smart speakers represent auditory IoT devices, while AI cameras embody visual modalities. Combining these creates multimodal IoT systems.
Modern smart ACs exemplify this: they use voice commands, computer vision for user tracking, and environmental sensors to optimize cooling. Such multidimensional perception is rapidly entering markets.
Many innovative technologies like projection keyboards struggle with adoption despite technical merits. AI product managers must align technology with market needs, creating viable solutions that deliver tangible value. Both technology and markets evolve dynamically, especially in AI.
Technological advances enable previously impossible scenarios. For instance, pre-2012 image recognition only handled simple tasks like license plate detection. Deep learning breakthroughs later enabled facial recognition for attendance and livestock management. By 2015, GANs revolutionized high-stakes applications like counterfeit detection and art authentication.
Thus, AI product managers need market insight, demand analysis skills, and understanding of model/algorithm principles to assess technical feasibility, optimize implementations, and balance resources—maximizing AI's product value.
They should grasp mainstream algorithms' capabilities and limitations without needing mathematical expertise, ensuring products deliver value even with imperfect algorithms.
(Content adapted from the author's presentation "The Transformation of Product Management in the AI Wave" at Jiangmen Venture Capital)