In the past two years, the artificial intelligence industry has experienced an explosive growth at home and abroad. The major giants have set up their own artificial intelligence laboratories and research institutes, but we have to admit that the current emphasis on artificial intelligence is mainly focused on technical types. In terms of talents, talents in artificial intelligence products have not received further attention.
The reason for investigating is mainly because artificial intelligence is still a very new type of technology. At present, there is no large-scale landing of commercial products for artificial intelligence. With the transformation from technology to commercial productization, artificial intelligence product managers will inevitably look like The Internet PM was valued many years ago.
What is a product manager?
Product managers are actually in every era, but they are just changing their titles. In the era of fast-moving consumer goods, because consumer goods have a standard, homogenization will be very serious. At this time, product managers are actually equivalent to marketing products. Managers sell homogenized consumer goods to users through marketing; in the software age, products are in a usable state, and they are rapidly occupying market share for product layout. At this time, the project manager is actually equivalent to the product manager. The main work is to manage the product development process for project management and quickly form a product landing. In the era of mobile Internet, technology has matured, and the competition between products is actually the competition of demand and experience. The product manager of the Internet has formed. A standardized system that taps the potential needs of users and enhances the user experience of products becomes the core competitiveness of a product manager.
Then we will have such a doubt, in the era of artificial intelligence, what is the core competitiveness of AI product managers? Is there a reasonable knowledge system to guide us to study?
Sorry, the answer is no! ! Just as the Internet product manager just appeared, there is no real set of knowledge and skill system for everyone to refer to. The current knowledge and skill system of Internet product managers are also colliding with the recruitment requirements of major companies and constantly in the process of product development. Exploring a common system, the current AI product manager, because there is no large-scale mature commercial products, the knowledge window that now faces can only be said to be more serious.
What skills should an AI product manager have?
I personally developed a habit. When I usually study a certain job skill, my first point is to search for the recruitment needs of various companies, especially giant companies like BAT.Often a new type of position is led by a giant company and gradually establish standards, we may wish to look at their recruitment needs.
Summarizing the recruitment needs of the above three giants, we can easily find that they have in common, AI product managers need to have the basic skills of Internet product managers, industry/scenarios and technologies are called the most frequently mentioned words, in terms of technology. It also mentions the need to have an understanding of the artificial intelligence foundation, which means that the technical manager does not require the product manager to reach the level of product development. This is similar to the requirements of the Internet product manager. Still in terms of scene mining and industry experience.
So we can basically get a conclusion,The core skill of AI product managers is actually to redefine the scenarios and requirements through artificial intelligence technology, and to redefine the way of solving problems through personal knowledge reserves, thus providing a feasible artificial intelligence solution.
What is the biggest difference between an AI product manager and an Internet product manager?
AI product managers and Internet product managers have a lot in common,Said AI product manager is in some respects a re-segmentation of Internet product managersAI product managers need to have their own unique skills in addition to ensuring that most of the Internet product manager's knowledge system, such as the understanding and mastery of artificial intelligence knowledge, and the definition of scene recurrence.
Then we have such a question, what is the biggest difference between AI product manager and Internet product manager? In my opinion, there are two main differences.
Difference 1: AI product manager's demand verification is more important
When we redefine the needs of the scene through artificial intelligence, demand and technical verification are the most important points. However, in the view of interconnected product managers, demand is something that is easier to find, and it is easy to solve it through conventional normal technical means. It is able to perform product iterations faster to form the product's landing and product closure.
However, for artificial intelligence product managers, this is even more difficult. Due to technical problems, we need to spend a lot of time to verify the scenario requirements, and verify whether the solution can be solved by artificial intelligence.For AI product managers, the availability of 60 points is often more important than developing a perfect product. Only when we can verify the scene through technology, is it possible for a product to be commercialized?Often, AI product managers will encounter a situation where they can easily find a scene requirement. In their opinion, artificial intelligence can greatly improve efficiency. However, after six months or even one year of technical verification, it is found that it cannot In this case, the need to move to other scenarios becomes even more important.
Difference 2: Technology iteration is faster
In fact, before the deep learning did not appear, the development of artificial intelligence has reached a bottleneck period. I once wrote a passage written by a person who studies artificial intelligence translation. The general meaning is that he read a paper about intelligent translation in Google. I found that all the previous technical accumulations have fallen behind, and I have to say that this is a huge blow. Therefore, if you are an artificial intelligence product manager, it is more important to pay attention to the cutting-edge industry technology update. Usually, one of the requirements verification last year did not pass, but it can be solved by technical means over time! !
Where is the main battlefield of the AI product manager?
For many comrades who want to transfer artificial intelligence products, they have no standardized guidance. I also consulted a lot of information and found that another ppt has reasonable predictions for AI product managers.
he didThe artificial intelligence products are divided into three categories, ten sub-categories, which are platform website classes. Most of them are open platform-like website products. Vertical scenes, this is a main battlefield field; the last is chat conversation class.Often this category is the most difficult, often combined with intelligent hardware, such as the recent hot smart speakers, pure platform-based secret products.
How do AI product managers learn and get started?
I personally think that AI is a skill-based profession. Its main opportunities are in the field of segmentation and cross-cutting. The biggest difficulty faced by AI product managers is actually how to define requirements based on scenarios.
The main learning route I personally think can be divided into three parts:
The first step is to find your own points of interest and specialties. It is best to have a field that overlaps with your own skills. In the case of technical categories, it is actually human-computer interaction/computer vision/natural language processing/biometric identification. These large categories are correspondingly divided into many small classes.
The second step is to choose whether your direction is based on the platform class, or whether the chat class is based on the scene class.
The last step is actually to implement the transformation. This is the most difficult step, and of course the final step. Here we focus on the implementation of the transformation step.
Implementing transformation requires us to deepen our understanding of AI technology. We need to cultivate a machine-like thinking mode and need to touch the multi-sensory human-computer interaction. In the design of artificial intelligence products, we cannot use the Internet to solve the demand. The means to apply, multi-sensory human-computer interaction is the biggest feature of artificial intelligence products.Of course, all products are actually going to be put into the scene. How to design a usable artificial intelligence product requires us to have profound scenarios and industry cognition capabilities, and product skills with cross-domain collaboration..
One point to note is that many of the inspirations in this article are based on some of Hanniman's views. He is the strategic officer of the Turing Robot, and he is grateful to him.
Finally, use his sentence as a summary:algorithmThe demo is more biased towards proposition writing, and when the product is commercialized, the key first step is to redefine the problem based on the scene! !
This article is transferred from the public number product Doggy,Original address