In the era of AI, the "change" and "unchanged" of product managers - "change".

In the past two years, large models and AI have been hot, and many industries and products given to AI will be reshuffled. What happens to the product manager in this case? In this article, let's discuss.

With the rapid development of artificial intelligence technology, we are entering an unprecedented era. Once, we thought that these high-tech technologies would only appear in science fiction movies, and I still remember the shock and envy that I was given when I watched the ghosts of Baker Street, the world of the game was so real, and the AI self-destruction was so touching, and I imagined that I could experience the life of Sherlock Holmes like Conan.

In the past, the application of AI technology often required deep expertise and high cost investment, but with the continuous and rapid development of AI technology, especially at the end of 2022, OpenAI launched ChatGPT, a general-purpose large language model, which has greatly lowered the threshold for enterprises to reach and apply AI technology. Enterprises no longer need a large team of data scientists or complex algorithms to easily integrate AI technology into their products and services.

The direct impact of this change is that more and more enterprises are beginning to explore and apply large language models to upgrade the user experience and the efficiency of the solution. From customer service to content creation, from data analysis to decision support, the application scenarios of AI technology are becoming more and more extensive, and many product managers are actively or passively turning to AI product managers.

So what are the new skills needed to move to AI product managers? To discuss this issue, we must first look at the types of companies in the AI era, and different types of companies have very different requirements for product managers.

1. Classification of companies in the AI era

In the era of AI, companies are divided into three main categories:

Category 1: AI R&D companies

Such companies mainly focus on the research and development of AI technologies, including algorithm innovation, model training, and optimization. For example, the now well-known OpenAI is such a company

The second category: AI service providers

The company provides AI-related services such as cloud computing, data analytics, consulting, and technical support. For example, Microsoft Azure.

The third category: AI application companies

This type of company applies AI technology to a specific industry or service to develop consumer-facing applications or industry solutions.

At present, most companies are in the third category or will soon become the third type of companies, that is, based on their own years of knowledge and experience in the industry/field, combined with artificial intelligence technology, to provide better products or services to customers/users. What we're talking about today is also based on these types of companies, which are the new skills that product managers need.

2. The "constant" and "change" of product managers in the AI era

1. The eternal goal of the product manager

If what product managers do in the Internet era is to ensure that the designed products meet the needs of users and are competitive in the market in a certain scenario. In the AI era, product managers need to combine AI technology to improve the existing technologies in the scenario, so as to upgrade the user experience and the efficiency of the solution.

For example:

If a product manager is compared to the director of a film, they not only need to be creative (product idea) and vision (film positioning), but also make sure that the film (product) can attract the audience (users) and be successful in the market (this success may be the box office or a breakthrough in a certain niche).

During the process, the director needs to coordinate team members such as screenwriters, actors, cinematographers, musicians, etc., to ensure that every aspect of the film is in line with the overall vision and goals. At the same time, the director also needs to consider feedback from different sources to improve the final version of the film.

In the era of AI, directors can obtain more creative inspiration through AI, and use AI technology in the production process to complete magnificent scenes that were previously impossible or require high costs to complete, so as to give the audience a better viewing experience.

The goal of a product manager is always the same

For the enterprise: Maximize the benefits (outcomes or impacts) we get from them with a minimum amount of development tasks (our inputs)

For users: Create maximum value for users and provide the best user experience

3. New skills needed by product managers

1) AI technology required by industry/productJust like the hot topic of discussion about whether product managers need to understand technology in the Internet era, the biggest advantage of product managers who understand technology is that they can communicate efficiently with developers, and similarly, a great advantage of understanding AI-related technologies is that they can communicate efficiently with algorithm engineers.

At the same time, mastering the application conditions and best practices of relevant technologies in the field where the product is located can help us know the technical achievability, how to maximize the value, and how to avoid the pitfalls that others have stepped on when designing the product.

Finally, when technical constraints are encountered, the plan can be discussed and adjusted in time to achieve the final goal. For example, if your company wants to introduce a large language model to upgrade an existing product or service, then we should at least know that:

• What is the relationship between artificial intelligence, machine learning, deep learning, generative AI, and language models, and understand the advantages and disadvantages of mainstream large models, and the capability boundaries
• 一些基本概念，比如预训练、督导式学习、自督导式学习、增强式学习，比如Token、CoT、Few-shot、Zero-shot、Instruction Learning、In-context Learning、RAG、Finetune等
• Some practical operations, such as prompting, how to prepare a test set, how to do algorithm filing, etc

2) Data analysis capabilities in the AI era

• Using large language models to improve the efficiency of data analysis, such as using large language models for sentiment analysis, helps us analyze user feedback more efficiently and understand whether users are satisfied with the product.
• Manual/automated labeling of bad cases, analysis of failure causes, and identification of common problems to iterate on the product

3) Human-computer interaction guidelines in the era of AI

• Basic research on intelligent human-machine design
• Human-AI- Interaction 等研究院的设计准则
• Augmented Reality (AR) and Virtual Reality (VR) in products

4) Security, Privacy, Ethics and Morals

• How to avoid AI from potentially saying harmful content, such as profanity, plagiarism, discrimination, excessive political correctness, and more
• How to avoid leaking personal data privacy, and how to achieve data compliance after introducing AI technology into products, especially multinational enterprises need to pay attention to both the "PIPL" and "GDPR"
• Understand basic ethical principles

Fourth, the end

Each of the above points hopes to provide you with some learning directions during the transformation, and I hope that the follow-up opportunity can give you a detailed description of the experience and pitfalls in the practice process of each small point. You can also share what skills you need to get when you transform into AI products? Welcome to discuss in the comment area~

最后我们想用一句很喜欢的话做结尾，All Product Managers are—or will soon be—AI Product Managers.

I hope everyone can transform into a successful product manager in the era of AI!

This article was originally published by @AI Practice Dry Goods on Everyone is a Product Manager. Reproduction without the permission of the author is prohibited

The title image is from Unsplash and is licensed under CC0

The views in this article only represent the author's own, everyone is a product manager, and the platform only provides information storage space services