Image source: Generated by Unbounded AI

We often talk about how to make a good AI product, but rarely talk about how to be a good AI product manager.

The well-known podcast Lenny Podcast recently interviewed Aman Khan, product director of Aman Khan. He has served as a product manager at Apple, Cruise, and Spotify, and is now building a platform for model capabilities and data detection. For AI product managers, there is enough to say.

The topics we talked about were very focused, how to become an AI product manager, how to become one of the 5% AI product managers, and how to find good ideas for products.

The article is based on the podcast, with slight additions and deletions from Founder Park.

Some interesting points:

There are three types of AI product managers: platform product managers, AI Native product managers, and AI+ product managers, but the core is to solve problems. The first step to becoming an AI Product Manager: Build your first product with AI. The AI ​​era actually needs product managers even more, because the hardest part is to know what to do, find the opportunities and problems that really need to be solved, and then clearly communicate what the AI ​​tool should do. Don't do what everyone else is doing. Just because the interface of ChatGPT seems intuitive to everyone, everyone likes to use it, and is familiar with it, it doesn't mean that this is what you should do. A good product manager must not only continue to create value for customers, but also make room for his or her team, continue trial and error, iterate, and accept the reality of rapid technological changes. AI is not just about automating problems, but making the user experience simpler and lowering the threshold for creation. As an AI PM, you must learn to "wander" in uncertainty, constantly testing and adjusting until you feel the product "drags" you forward in a certain direction.

01 There are three types of AI product managers, but the core is to solve problems

Lenny : Become an AI PM, we often hear this term in AI PM, just like becoming an AI PM through bootcamp certification. What's the easiest explanation for people to understand other than "PM working on AI"?

Aman Khan: The job of a product manager is to bring together design, engineering, business development, operations, sales and other parties to ultimately launch an impactful product. So, your responsibility is to represent the client and find a solution to the problem. In addition to these, I think AI can be divided into three types of product managers.

The first is AI platform product manager, which is my current role. This type of product manager primarily builds tools for AI engineers. For example, I am nowArize company job. We are an AI observability and assessment platform. We've been around for a few years and initially focused on machine learning, including ranking, regression, classification models, basically any recommendation system or black box that might be present in an application. Over time, we gradually expanded our business into the broader field of AI.

Now we mainly focus on large language models. There are many tools now built on the APIs provided by OpenAI or Anthropic, or on the interfaces of other large language models. However, these tools are still relatively cumbersome, the interfaces are also very new, and many functions are still being improved. After all, everyone has just begun to develop large language models, and the entire field is still in its infancy.

So, the current tools that can help developers understand whether their applications are working as expected are still very rudimentary and are still being developed. That's where I focus, helping AI engineers understand how their tools and applications actually work. If you are an AI product manager at another company responsible for impacting the company's business, you will know that you need to prove to your leadership that your application is indeed effective and achieves the desired results.

Next is the AI ​​product manager. The core product itself is designed around AI. Examples of this type of product manager can refer to ChatGPT, or the previously mentioned Raiza and NotebookLM, which are very typical AI products. The core experience here is driven by the underlying AI model, which can be said to be its "secret weapon." The task of these product managers is to package the technologies promoted by researchers and engineers into products suitable for businesses, consumers or users, so that they can actually apply these technologies. Simply put, their job is to package AI technology into consumable products.

The third type of product manager is the AI-empowered product manager I mentioned before. To put it simply, your original product manager job has AI technology to assist you, making your work more efficient. I think that as an AI-enabled product manager, you don’t need to build an AI model from scratch. Maybe you don’t have as strong resources and research teams as Google or OpenAI, but you can use existing large language models or Other types of models to build experiences that work best for customers. This is how AI technology helps product managers do their jobs better.

With the popularization of AI technology, the work of many product managers will gradually revolve around AI. In the future, AI technology may become a common infrastructure in SaaS applications like databases. So, if you really analyze the future trends, a lot of product managers will probably move towards becoming some kind of AI product manager, or developing for other product managers or companiesAI tools either build some cutting-edge AI models, like NotebookLM, or are based on some kind of Transformer model or GPT model to solve customer problems.

Lenny: What tools do you usually use to improve work efficiency and productivity?

Aman Khan: I am now super obsessed with these two tools, Cursor and Replit, and use them almost every day. They're great for prototyping and helping me figure out for myself what I can do. As a product manager, it’s very helpful to use these tools to create a working demonstration prototype. Of course, this prototype cannot be put directly into the production environment, but at least the buttons can be used, and the logic is almost built, and it can be used to tell a story.

There is also a great tool that many people may have heard of, which is Vercel, which is especially suitable for making landing pages. If you need to make a landing page, Vercel now has a starting template called v0. By inputting some tips, you can quickly generate a very beautiful landing page. And it’s not just the landing page. If you give it the right prompts and debug it repeatedly, you can even get a pretty good working interface. This is helpful when you and the design team discuss preliminary prototypes and interfaces.

If you are designing an interface, AI can help you do some graphic design. For example, you may need to design a logo, or you may want to tell a user story visually. Tools like Midjourney or DALL-E are great for helping you generate these visual assets quickly. Just learn how to get the results you want with the right tips. To me, it depends on the type of company you're in. If you are doing user experience related work, a tool like image generation may be particularly useful. For example, in 3D modeling, there are also some emerging tools being developed.

02 How to become an AI product manager: Use AI to make the product first

Lenny: When many people think of AI product managers, they probably think that what they want to do is products that deeply integrate AI. For those who want to change careers and become AI product managers, what advice would you give to those who want to enter this field, and what should they learn to enter a position in this direction?

Aman Khan: I am a bit of a layman myself. In fact, I studied mechanical engineering before, which has nothing to do with computer science. I learned many things on my own later. I haven't taken a course in machine learning or AI, and I don't have a PhD or master's degree. Coming back to the two types of AI product managers you just mentioned, one is to build the foundation for other AI engineers.A product manager for a facility and another is a product manager developing AI products. Regardless of the type, you have to focus on the end customer. Ultimately, you have to love the problem you are solving.

I remember you said something before, which touched me very much, that is, "If you fall in love with that problem, you will use technology to push its boundaries to solve the problem." For me, I have always been fascinated by helping very technical users, such as AI engineers and data scientists, solve some problems that I think are not particularly difficult, such as how to build a dashboard to understand how my model is running? It doesn't seem difficult, but as you get deeper, the complexity of the problem begins to appear. You come to realize that you're pushing the boundaries of technology, you're exploring possibilities and trying to move the field forward to solve customer problems.

The same goes for AI product managers. Your goal is to design the best customer experience and truly solve customer problems. If you really love this job and take the time to think about what is the ideal experience you want to solve, you will eventually find various tools to help you better understand and solve this problem. For example, let's say you work for a company and customers have a lot of support requests or complaints. Imagine the current situation of customer support. You usually need to enter a question through a chatbot, but sometimes the robot understands it wrong, or you have to find a lot of documents to solve the problem; or you describe the problem when you call, but you are not sure whether you are understood correctly. Long wait. If you are the product manager of this customer support experience, then the ideal product experience is that when customers ask questions, you can solve their problems quickly and efficiently. No matter what the process is, you only need to minimize these intermediate links. friction.

Assuming it started with a phone call, many emerging tools have begun to appear, one of which is very cutting-edge is the real-time API recently released by OpenAI. This API is a voice API that can have a real-time conversation with an AI-generated voice robot by providing some text. If you are a consumer, you can also experience it yourself. As long as you are a ChatGPT Plus subscriber, you can try this cutting-edge voice chat experience. Why is this interesting? Because if you, as a consumer, try this technology and explore the potential of AI technology, you may suddenly have a flash of inspiration and realize, "Wow, this thing may seem quite early stage now, but it has great potential." Then you can take this idea and discuss it with your team to see if you can use this technology to make some innovations.

So I think that’s the thing about this field: you need to be driven by curiosity. Your curiosity will take you in many different directions, helping you discover which tools can help you solve your problems, and even implement these tools in your own products. Overall, this is an opportunity to encourage you to explore what drives you and then find those tools that will truly enable you to achieve your goals.

Lenny: Digging deeper into this topic, assuming someone wants to be an AI product manager, what specific skills or technologies do you think they should learn to have a chance to enter this field? Should we specifically learn AI-related technologies, or is the most important thing about being a product manager still to make good products, and AI is only a part of it?

Aman Khan: Like you said, the real challenge is, what skills should you have to be qualified for this position? How do I make myself stand out? I think these are actually two different things.

First of all, you have to lay a solid foundation and understand the basic concepts of machine learning and AI. You need to know what is driving the advancement of this technology. Next, you should continue to delve into this direction according to your own interests. So, this actually has two levels: on the one hand, what basic knowledge do you need to know, and on the other hand, how do you apply this knowledge to the problems you want to solve or the industry you are interested in.

I think you should be driven by curiosity. For example, you can start with some basic learning resources, such as Andrej Karpathy’s basic introductory video on LLM, which lasts about an hour. After understanding it, you will be able to understand how this technology works and what its limitations are. You can also put this video into NotebookLM and maybe get a cleaner version. Another level is to try to practice these tools and try for yourself how to push the boundaries of technology. The two combined can help you improve your skill level.

In fact, if you look back at the early days, there were no specialized degree courses for product managers, let alone AI product managers. So, how do you set yourself apart from the crowd of competitors? You need to demonstrate an interest in the field, build a skill set around AI, and learn how to apply this knowledge to industries that interest you. This also answers the second question – how to stand out during the application process. Now is the time to catch the hiring manager’s attention by showing a portfolio of products you’ve worked on. Even if these products are just prototypes, showing them off can help impress hiring managers even before the interview.

To summarize, as someone who has been involved in recruiting AI product managers, the recruitment process actually focuses on three aspects: 1) Is this person qualified for the job I am recruiting? 2) How passionate is he/she about our work? 3) Do I enjoy working with him/her?

So, if you already have a foundation in AI and machine learning, and have some product portfolio, you’ve already answered two of the hiring manager’s questions before the interview. In this way, you actually save a lot of the interview process that requires you to evaluate whether you can do marketing, engineering, and core product managers one by one. My advice is to start with the basics, develop your own skills, and be driven by curiosity.Go ahead on your own and finally turn these skills into a portfolio that will stand out in your job search.

Lenny: I particularly like what you just said - in fact, it is not that there is a "shortcut" that can directly make you an AI product manager, but by building the products you have made, especially if you Products made with AI can make you look very valuable. For example, now using tools like Cursor, v0, and Replit, you can quickly conceive and make prototypes, and even draw sketches and do interface design through tools such as Midjourney and DALL-E. These methods are very practical. So if you want to apply for an AI product manager position but don’t have any completed products or prototypes, it’s definitely not easy to compete.

Aman Khan: In fact, it is easier to enter the field of AI product managers than before. In the past, you may need a more solid machine learning background to have a chance to enter companies that make AI products. Those companies then required you to have a deeper understanding of machine learning and models, including how data is prepared, how to partition training data, how to start a project, and the infrastructure involved.

But now, the role of the AI ​​product manager is more about building experiences around or for other AI product teams. For example, there are a lot of great videos out there that teach you how to build an app in an hour. I have a small example here, which is quite interesting: one time I was introducing Replit to a friend, and he said to me: "Now I don't have time to learn new tools, the technology field is changing too fast." I think the best The solution is to show him directly what is possible. So I turned on my phone, went to Replit’s website, typed in a prompt, and had it generate a website for me, and I could access it immediately. I asked it to help me generate a newsletter registration page. The first version did quite well, but I still want it to be optimized. So I re-entered the prompt and let it adjust the appearance and color. In less than five minutes, I created a complete registration page on my mobile phone. This was completely impossible a year ago.

I think that with the advancement of technology, these things will become easier in the future. As a product manager, your curiosity will continue to push you to explore new possibilities, and eventually you can become that Someone who always finds "cool" new things and brings them back to the company. So, that's part of the product manager's role: having a genuine curiosity about the space.

Lenny: About a month ago, I had a point of view. Now that there are more and more AI tools, many people think that product managers are useless. Why do we still need a product manager? Engineers can use these tools to do it themselves, designers can do it, and AI can automatically generate everything for you. So in this case, what’s the point of being a product manager?

Later I realized that it was exactly the opposite, precisely because of the productManager, you may not need other functions. What do you think are the strengths of AI tools? It's that they help you build things. You tell it what you want to do, and it does it. The hardest part is, instead, knowing what to do, finding those opportunities and problems that really need to be solved, and then clearly communicating to the AI ​​tool what to do. Next, you have to have taste and feel to know what a good product is and what will succeed in the market. Isn’t this exactly what a product manager does?

So, I’m curious what you think? Do you agree with this view? Is the product manager the role most likely to thrive in this AI era?

Aman Khan: I completely agree with this point of view. I think some executives or company VPs may feel that we have to rush into the field of AI, but they often have a kind of inertia and are used to doing things that have already worked. So I think the role of the AI ​​product manager in this kind of company actually represents how to maximize the use of these tools. If you already understand what these technologies can do, have a deep understanding of the customer's problems, and are the voice of the customer in your company, then you are the best person to figure out what should be done and how it should be done. This is a very strong position for the company.

Also, I would like to add to the core product skill you mentioned in the podcast - influence. These AI tools can greatly increase your influence, allowing you to more effectively communicate ideas to design and engineering teams, and even explain to company executives what should be done. So when you can communicate ideas to different people, these tools are really powerful in that regard. They allow AI product managers greater leverage. For me, and I totally agree, being an AI product manager is now the most influential position in a company, especially in this age of AI.

Lenny: Great, I'm glad we came to an agreement.

The point you just mentioned reminds me of Mihika (Figma’s product manager) whom we talked about before recording. Her super power is that she can be a design, engineering, and product manager. She is simply an all-rounder. Talent. This is why she is so successful at Figma, able to influence leadership to make products they were originally unsure about, such as "Do we really need to make a Slides product?" Now almost everyone can be like Mihika, both as a product manager and I can do both design and development.

03 How to become the top 5%: Don’t do what everyone else is doing

Lenny: What kind of AI product managers can distinguish themselves from ordinary AI product managers and become the top 5% in the industry?

Aman Khan: About two years ago, around this time, ChatGPT was launched.Everyone was impressed by it at the time, especially because its interface was super simple and intuitive, and the model's performance was so strong that it almost felt like human communication. Of course, the version back then was much simpler than today, and if you look back now, it's like looking at an older iPhone. At that time, you would indeed think, wow, this technology is very powerful, but if you use it now, you may feel that this thing is not perfect, and many functions are not powerful enough. Technology will continue to advance, and everyone’s expectations for it will continue to increase. However, that moment did feel like the “iPhone moment” for AI products. What happened next is interesting.

Then, many companies began to use this technology to try to build AI products. As a result, almost everyone's first product was another ChatGPT based on their own company's data. I don’t know why, but everyone is focused on this direction. Many AI product managers are saying, “We are building an internal chatbot based on our knowledge base that can answer questions.” But if you look at the actual use of these products , may not be as good as expected. I think it's worth pondering whether this is the right approach.

Back to this question, if you want to stand out among AI product managers and become the top 5% of talents, I think the most important thing is: don’t do what everyone else is doing. Just because the interface of ChatGPT seems intuitive to everyone, everyone likes to use it, and is familiar with it, it doesn't mean that this is what you should do. What you need to think about is, is this really what we need to do now? Is this a problem we need to solve? If you think deeply, you will find that the interface of AI can actually be very different. It may not be in the form of a traditional chatbot, but you may be optimizing or accelerating an existing process to help people complete tasks more easily.

Just like ourselves, within the company, we need to process a large amount of data to make decisions, so we try to automate this process. Although we also built a small chatbot to understand the technical principles, we made many different attempts in AI interface design. I think that many cutting-edge AI products are not exactly like chatbots. So, what you really have to ask yourself is, what is the most appropriate interface when designing this product?

Lenny: What I hear is that if what your team is doing is very similar to an existing basic model product, such as products related to AI basic models, that may be a red flag. , indicating that this may not be a big opportunity. So optimize for things that look different and innovative.

Aman Khan: Well, I think this is actually a bit like what everyone has been doing in the past two years, and now this moment has reappeared. The key point now is about the concept of "AI Agent". Let me mention again Yuriy (Grammarly’s head of growth)’s perspective. I listened to that show and found it particularly inspiring. He receives a large number of emails every day, and everyone says that they work in AI, AI Agent, this or that. But in fact, people often just apply this technology to a problem without truly describing the problem itself. What everyone is talking about is a solution—for example, an AI Agent to do something.

So if you really analyze it, you should ask yourself whether it is really necessary to build an AI Agent within the company, or should you let those basic model companies develop this agent layer, and then your job is Let this experience be seamlessly integrated into your existing products so that users don’t feel that it is AI at all. I think this is the most powerful place and where product managers can truly innovate. Let’s be honest, it’s unlikely that you will have the capabilities of top researchers to design a new model from scratch, or devise some new technology to build the next agent framework. But what you can do is take existing technology and find a way to apply it to your company, and that's where I think the AI ​​product manager role stands out the most.

Lenny: I guess hearing this now, a lot of product owners are thinking, "Great, I gotta get everyone on the team to hear this, because I guess every product manager and every Companies are looking at these AI ideas and keep bringing up various AI-related ideas. Everyone thinks, "Oh, we have to do AI." So there are countless AIs. Product creativity.” So I think this advice is really important, and many times everyone has reminded us to focus on solving problems, rather than just doing AI for the sake of “doing AI”. AI is just a tool to solve a problem, not the goal itself.

Lenny: That’s very well said. What is the difference between AI PMs in general and the top 5% PMs? Do you have any other thoughts to share?

Aman Khan: Actually, I want to steal some advice from our friend Kevin Yien (Stripe product manager). He mentioned it in your podcast. I think it resonates particularly well with me personally, that is "You have to be able to walk and chew gum at the same time."

As an AI PM, your job is not to make AI products, but to solve customer problems. I think this core idea is important, and we keep coming back to this topic. You will be constantly pushed by the team and the company to solve a certain business metric and drive a certain KPI. But your real goal is to make room for other things, like prototyping, trying out tools yourself, giving the team some time to do hackathons, trying out tools themselves, and tearing apart product experiences that work really well. For example, our company did a live broadcast of the disassembly of the NotebookLM product yesterday. We will select a cutting-edge AI product and hold a webinar about every two weeks.Analyze how it is done. These three activities don’t actually directly affect your KPIs or business metrics, but they help us better understand this area.

So I go back to my previous point: you have to be able to walk and chew gum at the same time. Not only must you continue to create value for your customers, but you must also make room for your team to continue trial and error and iterate. You’ll find that some experiments may not inspire users as much as you expected, but you’ll learn a lot in the process. This is also a strong point. To stand out in the field of AI PM, you have to accept the reality of rapid technological change. You may think that a certain technology will bring great results, or the product you make has a poor experience. Like you mentioned before, good cooks occasionally make a subpar meal. The same goes for AI products, you will experience some failures, but the goal is to keep iterating.

This is how AI gets better, and this is how companies can better deploy AI. So don’t give up on your original direction, but find ways to make AI scalable within your company. I think a really good AI PM would be able to do just that.

Lenny: I feel like the metaphor of "walking and chewing gum" doesn't quite convey the difficulty that Kevin is trying to express. If you think about it, walking and chewing gum don't seem that difficult. Maybe there is a more appropriate metaphor.

Aman Khan: Yes, I don’t quite understand why I use this metaphor, but it has always been used to describe doing something very difficult, and it feels like you can’t do so much at all.

Lenny: Haha, indeed, we don’t dwell on this metaphor. Is there anything else like this that you think would be valuable to share or discuss, especially about how to be a successful AI product manager that we haven’t covered yet?

Aman Khan: The organization might want to do one thing and you're trying to get them to do another thing, and that's exactly the challenge in this kind of role because there's so much noise and excitement around AI , knowing to build the right thing is exactly what this job needs to convey. So I think when you put all those signals together and try to solve a customer problem, that's ultimately what it really means to be a successful AI product manager. This is the most difficult part of the job, but also the most important and impactful.

04 How to have good ideas: work hard on user experience

Lenny: How do you find some really interesting and promising AI application ideas in your company? Is it through a hackathon? Or set some goals to ensure there is an AI element in the solution? Are there any practical practices that can help you find these good ideas in the direction of AI?

AmanKhan: There are actually three methods in the company. If you start trying it tomorrow, you should see some good ideas emerge.

The first one is what you just mentioned, how to measure the effect of AI? I think it’s easy for product managers to focus on metrics, and in fact, I think every AI product manager needs to have a metric. The PM in every company has a task to promote a certain business indicator, but in fact, no one seems to have a clear measurement standard for the development of AI prototypes, which is actually quite interesting. I've talked to a lot of companies and asked them how they measure the impact of AI in their business, and their answer is, "Oh, we don't expect it to directly lead to revenue growth." So how do you know if what you are doing is effective? So, I think you need a metric, which is how many times you try in this regard.

This is also related to the second point, which is the hackathon method. I think hackathons are particularly good, allowing everyone to experience technology first-hand, and get rid of the feeling that "AI is not easy to use." The goal is to get everyone in the company to try using this thing, and I think this is also a good opportunity for AI product managers. Many times, people are a little resistant to hackathons, thinking that they take up the company's time, but in fact, you can go with some problems to be solved and see if they can be solved using AI. You might end up with 10 questions, 9 of which you try to do with AI and it doesn’t work. It's like, the AI ​​doesn't work as well as you thought it would, or you want a very specific answer, but the AI ​​doesn't do that well at those questions.

I have an example. We did a hackathon ourselves a few weeks ago, and the engineering team came up with various ideas. One of them was to make a Slackbot to help send reminders to people on duty. We originally thought this was very simple, that is, if someone said they encountered a problem in the support channel, Slackbot could automatically identify and push the problem to the relevant person. We thought, isn’t this the perfect use case for AI? Let it analyze the problem categories and then pass the information to the right people. It turns out that this is actually a difficult problem to solve because there is a lot of context missing from the information. For example, this person worked on this problem two weeks ago, but now he is working on other projects and another team has taken over this part of the work. So, the key is to find the right problems to solve. A good intuition is to use hackathons to identify problems and see which ideas work and which don't.

Finally, I think the most important point is to think deeply about the user experience. You have to find problems to solve with AI, but in fact, you can also look at existing successful AI products and see what they have done right. For example, many people are trying to use AI to completely replace certain jobs. There is an interesting story that Betty Crocker (pasta brand) once launched a cake mix that only requires adding water.You can make a cake. As a result, the sales of this product were dismal. Later, they changed it to require adding water and eggs, and sales suddenly surged. Because they realize that customers actually want to have a sense of control over their experience. If you look at current autonomous driving, although it is fully autonomous, you can still control settings such as air conditioning and music. Customers still want to have some control over their experience.

This is how something like the “IKEA effect” manifests itself in AI products: when users feel they have an impact on the final experience, they will feel more involved, rather than having everything automated. This is an important lesson we have learned from some successful AI products, and perhaps you will find more similar examples in this field.

Lenny: This is such an important point. Although you can make a fully automated product, you don’t necessarily have to do so. And the Betty Crocker example is also quite interesting. It reminds me of Blue Apron (the food delivery company) and those cooking services. A friend of mine recently told me that he cooks so much that he finds time to cook, only to find out after asking that he uses Blue Apron. On the surface, it looks like he is cooking. You can also say to others: "I have been doing a lot of cooking recently," but in fact it feels different.

Aman Khan: Yes, really. I want to dig a little deeper into the Blue Apron issue. This is a very interesting point. Look, what does Blue Apron do? Are they trying to keep people fed, or just filling their bellies? I don't think so. If you really want to solve the problem of eating, you might build a product like DoorDash (a food delivery company), where food is delivered at the touch of a button. But the goal of Blue Apron is actually to bring people closer to the experience of cooking, or the feeling of "cooking".

I think this is also what great AI products do. AI is not just about automating problems, but it makes the user experience simpler and lowers the threshold for creation. So the role of AI is not to automate everything, but to make it easier for users to use it and create a better experience.

05 A good product manager should drive the project with energy

Lenny: You have consciously chosen to be an individual contributor. -contributor PM*), rather than going to a high-level position as a product manager, such as being promoted to director, VP, etc. Success and long-term success in an IC role require specific skills and mindsets. Many people either can’t go up any further, don’t want to go up, or they haven’t even considered the possibility. So my question is, what habits, mindsets, and lessons learned do you think are particularly important to long-term success as an individual contributor??

Individual-contributor PM: A PM who does not assume team management responsibilities and focuses on personal projects.

Aman Khan: Looking back on my own career, many things actually depend on your personal driving force. I am particularly obsessed with solving customer problems. In order to do this, I may have to spend a lot of time delving into the details to understand the essence of the problem. So I think there are three main aspects that need to be paid special attention to in order to be a successful IC PM.

First, I want to set the tone a little bit, which is that being an IC PM is really hard. I mentioned before that it becomes easier to get into the product manager space to a certain extent if you have access to these tools, now that you have these tools to build the next prototype, but I think the standard and The impact needed has been raised. So I think the challenge is, you have a really tough job and the signals are constantly there and things are constantly changing, how do you really stick with it? So I think there are three things to consider here, which are energy, waiting and wandering, and amplifying signals to make decisions.

The first is energy. I think that's one of the biggest lessons I've learned working with our CEO: energy is really important when you're not sure which direction to go. If you go into a meeting with more energy and enthusiasm, you'll find that many obstacles can be overcome. Conversely, if you feel down a bit or aren't interested in an idea, everyone will feel it. After all we are dealing with people. If you change your mindset and bring a little more energy, many times you'll find that the conversation flows more smoothly and feels like a conversation with a friend rather than a difficult discussion. In fact, you may also face uncertain situations, find it difficult to make decisions, and be in a bottleneck period. In fact, this is something that everyone will encounter, like the "bottleneck" of the product, not the "bottleneck" of writing. You may be waiting for inspiration for your next article, but as a product manager, you are waiting for an idea or goal that will move the project forward.

Let me give you an example. Almost a year ago we had a difficult choice: should we invest in LLM and large language model technology, or should we continue to focus on existing customer groups. This issue cannot be decided clearly at the moment. So I did one thing with energy, which was to contact all the people on LinkedIn with the title of "AI" and start a phone call. Through this process, I not only learned how to better communicate our problem and product, but also sent a signal to the team that I was willing to dig into the details with everyone, even while busy advancing existing projects. Among these "chores". I think this attitude of leading a team is very important. When everyone is confused about the direction, you can appear on the court as a "player coach" and promote with a positive attitude and energy, which can greatly improve the overall efficiency of the team.

So I think, likeLike LeBron James, you're not just the coach, you're personally involved in every piece of the puzzle. This kind of leadership is also an inspiration given to us by our CEO. Although he would be tired, I could feel it, but he never showed it when he came into each meeting, always with energy and enthusiasm.

Lenny: I heard you share two points that sound great. The first is that being positive and energized during a meeting can really bring real energy. The second point is to actually do things and solve problems, such as taking the initiative to enter sales, maybe becoming a salesperson, contacting customers yourself, and doing some "sneaky" things. Regarding the first point, how can we bring energy? Is it simply "loud"? Or just act energetic?

Aman Khan: I think it depends on the individual. This is a very personal thing. For example, for me, I might show excitement and participate with my own energy, but for others, it might be just small things like caring about how others are doing and trying to keep the atmosphere of the meeting positive. I think everyone feels the energy differently. In fact, this is also related to a point I mentioned before. Human beings are very good at picking up other people's subtle signals. You need to convey an energy that makes people feel that you are committed and 100% involved. So I think the key is to make everyone feel that you can increase the energy of the entire team instead of decreasing it.

Lenny: I have worked with a project manager before, and he completely fits this point. Every time he had a meeting, he would come in with a problem-solving attitude and be super positive and energetic, which would change the entire feel of the meeting in one fell swoop.

Aman Khan: I remember a little story in the early days of Lyft. I have some friends who were at Lyft in the early days. Their self-driving team had a tradition. At the end of every meeting, they would say "make it happen” (let it happen). It sounds simple, but it's super powerful. It's a "we're going to get this done" attitude. This has even become a culture for them. After we make a decision after the meeting, we must implement it.

This small habit can feed back into the team culture and motivate the team. You actually have the opportunity to shape the culture of the people around you and influence the atmosphere of the team, especially in the role of PM. Everyone looks to the PM to make decisions and solve problems. Even if it is a difficult decision, you can make it. Energy that affects the entire team.

Lenny: Totally agree. Okay, first of all, your advice is: energy is really important. If something is not going well, bring energy to the team and not just keep people positive in meetings, but also do the work yourself and get to the heart of the problem. People will see the time and effort you put into it, which can help solve the problem. question.

Aman Khan: Yes, we can even continue from the second point. I think there is aThe important point is that when you work with others, you will have more empathy for their work. Because they will tell you what is difficult and challenging about the job.

I have found a consensus in many teams that no work is unimportant. So as a PM, if you work with someone to solve a problem, you're actually learning from their work and understanding their struggles, and you're also pushing the entire team to move forward better.

06 Learn to "wander" in uncertainty

Lenny: Go ahead and talk about some other experiences you found helpful in your IC PM role. You mentioned the two concepts of "waiting" and "wandering". Maybe we can talk about this.

Aman Khan: In fact, the concepts of "waiting" and "wandering" are really very powerful. I feel like it's very close to my own heart, especially when I don't know what to do next. We talked about energy before, how to bring energy, but that doesn’t solve the problem of what we should do. As a PM, your job is to figure out where things are going, what problems need to be solved, and how to actually solve problems for customers.

So you have to think "zoom in" from time to time. It feels like you are in a camp and need to decide who should lead the team and guide everyone forward. In many companies, especially large companies, there is often a tendency: wait and see before doing familiar things. In the field of AI, there will be such thinking: "Should we wait and see how things develop, or should we take the initiative to guide the team?". Many companies will choose to wait and see if new technologies or models emerge before making a decision.

For example, everyone is saying now, "Let's wait for the next version of OpenAI to come out and see what products we make." I have heard a lot of similar statements. But I think that sometimes it is reasonable to wait, especially when the technology has not yet fully developed. But at the same time, there is also an important part of "wandering", that is, you have to take the initiative to find the direction and think about the next step. How to go. To me, the role of the PM is that of the “wanderer.”

I remember one experience at a company where an engineering manager and I were chatting about this issue over beers. He said that sometimes he feels that he is not 100% dedicated to the work, especially towards the end of the work, when some things are uncertain and he does not know where to go. I found it very interesting to listen to what he shared at the time. In his opinion, the team was actually in good shape, everyone’s morale was high, and the plans were clearly arranged. But from a product perspective, we are very confused and don’t know what to do next. As a PM, you have to take the responsibility of leading everyone to find the direction. This process is not easy, and sometimesIt's like groping in a fog.

At that time, when I was discussing with the founders of the company, Aparna (our chief product officer) once told me: "This is the process from zero to one. Sometimes you are really not sure about the next step. How to go, but you know, you will eventually find the right direction. "So, as a PM, you must learn to "wander" in uncertainty, keep testing and adjusting until you feel the product "drag" in a certain direction. ”As you move forward. Therefore, it is important to adapt to this uncertain state and "wander" rather than "wait". While others may be waiting to see how the situation develops, you, as the PM, may need to take the initiative to explore and lead the team.

Lenny: What I take away from this is that if you don’t have a clue about the future of your product right now, that’s okay. Anyway, as you keep trying, you will eventually find your own way. Moreover, making the prototyping and ideation process easier with the help of AI tools theoretically reduces your stress because you can try different solutions.

Aman Khan: I think AI is like amplifying signals through noise. Signals are sometimes sparse and the information you can get is limited. But you can use some powerful tools like we use Gong to analyze what potential customers or engineers are saying and see where their pain points are. I can't go into 100 meetings a week, but I can put the minutes in there and feed those into some language model that can handle large chunks of text, like we did. This way, you can identify the most common problems. Discussions around these issues can also be very effective.

Now you can also use your voice to process this information and integrate everything. You can even make it a link in NotebookLM. The key is to find ways to extract the signal from the noise. This is equivalent to giving you a "superpower" because you can focus on multiple places at the same time and use technology to help you find valuable information. So, I suggest you use AI to improve yourself and catch those signals.

07 Enjoying the process will take you further

< /p>

Lenny: Is there anything else you can share about your success?

Aman Khan: Being a product manager sometimes feels very stressful, and the decisions you make often have great impact. But actually I remember having a discussion with one of our board members, who was a multiple-time entrepreneur who had taken companies public. I asked him what advice he had for an early-stage product manager. His feedback surprised me. What he said was: enjoy the process. In fact, I think as long as you stay curious, learn, and make products for customers you really care about, this kind of driving force will make you go further.Far.

I personally feel that it is more interesting to maintain a high-energy state, to truly care about what you are doing, and to be more motivated to do it. If you keep learning and having fun, you'll iterate faster too. So, my advice is to enjoy the process, really just enjoy the journey. Even if you are confused sometimes, enjoy the process.

Lenny: I really like what you said. I feel the same way!

Recently I heard another suggestion about public speaking, which is also related to "enjoyment", which is to participate in a course called "Ultraspeaking". One of the key points in it is: the speech should be interesting. Even if you feel nervous, rethinking it and figuring out how to make it fun can really work.

Aman Khan: It's more fun than no fun at all.

Lenny: Even in the face of pressure and fear, if you think about it from another perspective, how can you make it interesting? Most of the time, it's not the end of the world, or even the end of my career. So, try to enjoy it.

Aman Khan: Totally agree.

Lenny: Let’s jump into today’s quick Q&A session. The first question is, what are the two or three books you most recommend?

Aman Khan: I particularly like Bill Bryson’s A Brief History of Everything. This book was interesting to me because I have always been interested in science and read a lot of novels as a child. The book is about the history of science, especially how we know these things. For example, you may find that Newton was a little unfriendly to his colleagues, or that Darwin was not so optimistic at the time. Many scientists' discoveries were not immediately recognized by everyone, and it often took several years before anyone realized the importance of these discoveries. This book also gave me a better understanding of the scale of the universe and the history of humanity.

I think it makes you rethink time and space, especially now that we live in the age of AI and that all seems trivial. The stories of scientists in the book are also quite interesting, so I recommend almost everyone to read this book because it is an easy read.

Another book is "Life Design Lessons", which was written by one of the founders of IDEO U and is also the work of professors at the Stanford University Design Lab. It is very practical and contains many exercises to help you clarify the direction of your life. An exercise I did at the time was to write an article about the meaning of work, then write an article about the meaning of life, and finally find the intersection of the two. This exercise made me really realize what my goals are, and is especially suitable for those who are career-confused. In short, these two books help you look at life from different perspectives.

Lenny: Are there any newly discovered products that you particularly like recently? It may be an AI product or it may not be.

Aman Khan: I think we just talked about some AI products, such as Replit and v0, are super easy-to-use prototyping tools. I think they can ultimately produce very exquisite products. However, one product I've been having a lot of fun with recently is Websim. What’s special about this tool is that it allows you to prototype really interesting, even crazy ideas that you might not have even thought of. It’s hard to describe it in words. I remember I saw the speech of the founder of Websim at an AI gathering before. When they were demonstrating the product, they went directly to Reddit and looked at the Websim subreddit. There were many posts there, and they randomly picked a few. , see what everyone is talking about, and try to describe that content. I've been really obsessed with it lately, it feels like it pushes you to break out of your routine and try to do things that you didn't think were possible before.

Also, I have another product. Actually, I'm a bit old-school and still prefer to write in paper notebooks. I'm always looking for the best notebook to use, and recently at an AI meetup, someone gave me a notebook. It is made of recycled apples. When you open it, you will find it smells like apples! This really surprised me because I didn't think it would have that much of an effect on me, but that scent actually made me feel a positive feeling every time I opened the notebook. It's like an instant sensory experience. When I smell the fragrance, my mood gets better, which makes me enjoy using it. I don't know if I will buy another notebook when I'm done with it, but I do like it now and the little scent makes me feel particularly interesting.

Lenny: Do you have any life mottos that you often use to motivate yourself, or that you find particularly useful in work and life?

Aman Khan: Yes. Steve Jobs said: "Your time is limited, so don't waste it living in other people's lives." This sentence touched me a lot, because many times, we will feel professional pressure, You may feel like you have to find your next job quickly, or start a business and do something crazy. And I think what is truly powerful is being able to re-examine what kind of life you want, rather than catering to other people's expectations of your life. So I always keep this sentence in my mind. Whenever I am confused, I will think: What kind of life do I want to live? It's best to think in terms of this when making a decision.