#114 AI PMs Will Take Over Your Job - Ofir Natan @PracticalAI
In this episode, Alex and Christian are joined by the insightful Ofir, a Chief Product Officer with a deep passion for AI. They explore the evolving role of AI in product management and design, discussing how AI capabilities can empower users, boost productivity, and unlock new value. From AI copilot assistants to generative AI tools, they provide practical examples and valuable resources for integrating AI into your work. Join us as we dive into the AI revolution and discover how to develop an AI mindset in the world of product development.
Connect with Ofir:
Link to Ofir's AI talk:
Link to the resources:
What We’ve Covered:
(00:01:24) A good CPO focuses on AI, tools, and AI mindset.
(00:06:59) AI empowers product managers to innovate faster.
(00:09:13) English is the new programming language.
(00:13:50) Will AI make humans smarter or dumber?
(00:17:41) Rate of change challenges the education system now.
(00:21:50) Generative AI tools will help.
(00:24:45) Saving time with backlog management.
(00:30:06) Revolutionary shift in product development approaches.
(00:32:32) AI transformation empowers users and boosts productivity.
(00:35:00) AI is a powerful partner, not hype.
(00:39:23) Be mindful of AI but don't underestimate it.
Meetings Notes (Transcript):
Christian:
So, another great guest at the Product Bakery today. I am more than happy to welcome OFY and Nathan. I think it's fair to share why we're here today because Alex and myself as well as you, OFI, we've met at the Chief Product Officer conference a couple of months ago in Berlin. And Ofyo, you were holding an amazing AI talk that blew me away, at least Alex, I guess was the same for you. And we thought there are so many cool, crazy, practical things that you are sharing that it's worth having a conversation today. And yeah, that's why we actually here. Mean, as always, I like to do a quick introduction because Nathan, you've been around for quite a while in the whole world of software as well as AI, and you used to work for many years, more than a decade, as an software engineer. You used to be a chief product officer.
Chief Technology Officer. And you said at the beginning that you are now having your own big project with a lot of power, but it's not manpower, it's actually machine power because you are so focused working with AI and helping companies to make those transitions. And I just want to share before we dive into the topic a little bit on how we usually record a podcast episode, because for our audience it might be interesting because we do a short pre call where we talk about the topic, set the frames, roughly a frame on what we want to talk about. And ofYA, you said that there are a couple of things you want to give away to our audience today. And I think it's worth highlighting that because I believe that our audience loves the fact that we work with many examples and try to make the deep dive. And you said that you want to talk about examples on how to use AI and tools to, for example, create product content like stories or prompts and stuff like that. And at the same time, you would also like to share with us a lot of links to tools that people can use. And two more things are you want to help people to develop an AI mindset
So that's something that we definitely going to focus on. And due to the fact that this is an audio recording and the whole AI stuff can be also very visually. We are going to add to the podcast description a lot of links as well as a talk from Nathan so that you can do from Opiate, sorry that you can do a deep dive again and see actually what we're talking about today. So what you as a listener need to do is you can just relax. You will be able to find all the resources in the podcast description as well as a link to Opia's LinkedIn page where he is regularly posting updates. So with that said, I think a way I would like to kick off this big, big topic of AI is a question which is is the AI revolution coming or is it already there?
Ofir:
Yeah, so that's a great question and I think I can give examples of the phone and the mobile phone evolution over the physical phone. If you were to ask someone back in the day where the phone was connected to the wall, is the communication revolution already here or is it going to happen? The reality is they probably say, yeah, this is the cutting edge. And so I think we're just in the beginning where machines don't really understand us, they just predict what is the next token or word or text or code piece would be. Basically they've been trained on data and they just output their predictions for the next thing that they think they predict is to be the right thing. But I think the barrier of artificial general intelligence or that point is ahead and we're going to see a lot of interesting things happening including having the machines walk around between us and become integral part of our lives. And it will be very hard to distinguish between what is AI generated and what isn't. So I think we're just in the beginning.
Alex:
I mean, hearing you say that makes me think of like should I be scared of the future where AI is like walking with us and us not even realizing or being able to tell them apart from what is real?
Ofir:
Yeah, I'm thinking the same. But I do think that like any other disruptive technology, like mobile and cloud, this is something that will essentially be the evolution of work, of lives and it's not going to be as scary as it probably looks like from where we are right now. And I do think that what is different about this revolution is how accessible and easy and it's becoming even easier to utilize those tools and capabilities and improve our lives. Sam Alfman, the CEO of OpenAI said that he thinks this would be the greatest revolution and empowerment for the human race, the biggest platform shift. So I believe this is something that will benefit us more than annihilate us, I hope.
Alex:
I think the way I like to look at it is it's a really good partner or it can really help us become a better version of ourselves. Right, because as you say, it just equips you with all that knowledge. I mean, I can suddenly be a great copywriter and I am terrible at writing, but I can write a campaign or I can write a blog post and so on. And it gives me this additional skills that I otherwise would not have or that would take me another life to learn. Right. If I look at all the copywriters that are probably fearing for their jobs now, I think it's really good to have these tools as an addition to what you already can do and to expand your overall skill set.
Ofir:
Yes, I agree. And I think this is definitely I like to say that AI will not replace product managers, but product managers who master AI will replace those who don't. So it will basically make us more full stack and more capable and faster and more productive. And we can iterate very quickly. Things that today block us from moving and experimenting quickly, which is velocity is probably one of the biggest aspirations or goals of a product team today, or a company, a product company, they want to move fast and this equips us with the ability to move faster and experiment quickly and then improve on our products. So I do think it will benefit both product teams and both consumers by getting more value very quickly. And we can already see this. The rate of new releases and new features coming out every week is pretty crazy.
And the reason for that is also that this technology kind of builds on top of existing technology and the iterations are becoming faster and more exponential. And this is something that this is one of the reasons why I really recommend product teams and generally companies and entrepreneurs, to really experiment very early because it's going to be hard to predict when that exponential moonshot will happen and the companies that are already working on it or making it accessible to their customers or adopting these tools internally will just surf that wave and leave the rest in the dust. And that's something that we don't want to miss.
Christian:
Yeah, but let's talk about it because you mentioned that product managers who will not work with AI will be replaced by the people who will. So how do you see the whole product management product manager role and also product design role going to shift in the future and also what are the risks for those people?
Ofir:
Yes. So I'll start with maybe our listeners can imagine a world where someone without coding experience or design skills can will into existence no working application and digital experience using simple language. Once you understand that English is a new programming language and the new design language. And I'm just saying English is natural language, right? It understands all languages, that world is practically here or at least partially here. And that changes the role of the future product manager and designer. First of all, the product managers of the future will I'll start with having almost every product will have a digital assistant or digital experience that is AIpowered. So if today product managers understand cloud technologies, they speak fluent SaaS and they can communicate with their entire teams around it. If every product has an AI component to it, then every product manager is essentially an AI product manager in some of its responsibilities.
And we're going to see more full stack professionals right. You say, I don't know how to write content, I'm not very good at it, but now I can actually come up with reasonable outcomes. Then we'll probably see a shift into more of the editor versus the creator. Because if you can generate an image, you can only edit on top of that and change it and modify it. You don't have to give birth to a new story design product. You can just iterate on top of something that is half baked and that saves a lot of time. So we'll see boost in productivity, quicker iterations, more data informed and user focused product teams. That's what I predict.
And yeah, this is already happening today.
Christian:
But does it mean that product managers become more full stack? Or could we rather say that all product managers will become great entrepreneurs and quit their jobs because they know how to build their own stuff?
Ofir:
I think it's a good question myself. I did that exactly. But I think that entrepreneurship is kind of this internal driver that not everyone really wants to be an entrepreneur. But I definitely think that entrepreneurs in spirit that were kind of caged within the boundaries of what they thought they could achieve or do will now get inspired to really build things and developers that are entrepreneurial in spirit will be able to launch full products in ways that they've never been able to do.
Alex:
I mean, in general, I think it will just make it easier to bootstrap a company. Right, independently of your background. I feel like in my whole career, I've seen a lot of colleagues of mine who suddenly started something, most of them were engineers because they had the advantage of I can build something and get it out to the market or other people who learned coding. But I feel like now this gives every person the tool. Like is it someone with a business background? Is it someone with a design background? Is it someone with whatever other creative background who has a great idea to get a first prototype out of the door without going all the way back and starting from scratch. A new universe.
Ofir:
I think that's what Sam Elkman said that it's the greatest empowerment tool, I think it will democratize skills in a way, because the hard skills or the kind of the technical skills would not be as important as the business idea and the execution and the improvements in iterations and customer focus that would be probably the drivers. And the execution part would be not as much of an issue, at least in the early stages.
Alex:
But that brings me a bit to another slightly more controversial question, right? Like, will it help humanity actually to become smarter or will it just make everyone damper? Right? I mean, it's the same with like since we have Google Maps and Navis in the cities, I don't know how to get from point A to B because I never did it without. Right. And I think also the human brain is used to then always go for the more comfortable route. Will it end up that there is people who have the understanding and who have the knowledge about architecture and so on, that then are the rare kinds and principles that will stick around in the companies because everyone else won't learn it because they have the tools at hand to just automate it for them?
Ofir:
It's a good question. I think we can probably look at it as a good example would be the calculator. It was a big fear amongst kind of accountants and analysts where they did everything manually on paper and when the calculator came in and then everybody said, oh no, it's going to take our jobs. And eventually they just adopted this. And the skill of doing math in your brain just became obsolete the same way as your memory of phone numbers becomes irrelevant when you offload it to a machine. So I think that it's an evolution. It's not going to make us dumber. It's just going to make us really more efficient and being able to focus on things that matter.
The CEO of Microsoft said it's going to take the drudgery out of work and life. Right. Those mundane tasks, those things like searching something online takes time, right. Researching, reading a lot of papers just to get the bottom line, or a statistic, that's something that takes more time and not really makes you necessarily smarter, essentially if you need it for a specific task. And I think we will just be more focused on things that are more of interest to us than things we need to get done in order to move on to a certain point. So I think it's an evolution, really.
Christian:
Yeah. And I think that brings us also to the tools and the available options we have right now. But before we go into that, there's also like a controversial question from my side because if we look back at the industrial revolution, the human race and people had like decades and actually a century of more than a century of time to develop together with that revolution, right? So people were able to adjust their jobs and stuff like that. But if I look at the AI stuff, let's look forward one or two years. Customers support jobs, they'll be gone. Content creation will be mostly gone. Coding in the future will be gone. So what I see is that many people will lose their jobs, first of all.
But at the same time, it's also something we discussed before we started recording is that we will see much more specialized roles. So people who will be the one who as you said at the beginning, right, mastering the tools in terms of coding perspective or product perspective or design perspective. So how do you see that? Am I too critical and too negative about it? Or is there also this risk of being more efficient, but with a big cost of many people losing their jobs.
Ofir:
It's a good question. I'm thinking the rate is really the pace of things, the pace of change is really a challenge here. But right now, I don't think it will happen within two years at least. Okay, no yet. Right? But I don't think it would take two years. But I do think people will change and adapt quickly because they must. And companies will probably design new experiences on top of AI generated or AIpowered experiences and products and services, and we'll just see a new wave of possibilities. So for Was, I attended an Education Technology AI conference from the Ministry of Education, where I was thinking, oh my God, can education school system keep up with the rate of change? And I was blown away by what teachers and students are doing actually already now.
And they did a survey here in Israel. Almost 70% of the students think that AI will benefit them more than actually harm them and teachers the same way. And they're already using it inside lessons. So the kids and the younger generation, and they're adapting very quickly. And I do think that most people are capable of changing with the time and moving on to the next value adding role that they can take on. But I do think governments, corporations, enterprises, this is the responsibility also of product leaders and team leaders to make sure that their teams are upskilling and experimenting early so that it doesn't catch us out of nowhere, kind of, hey, I didn't see that one coming. We need to be on top of things. And I think, as I said to the product executives in the talk where we met in Berlin, it is our jobs to really experiment early and empower our teams to experiment.
I do hope that this is also coming from the top to bottom, and not just from the people using cha, GPT and midjourney and so on, but also from executives and management that they're seeing taking responsibility for their workforce, for their people and adopting a human centric strategy and not, hey, let's cut all the jobs and do that and save money. Right?
Alex:
No, absolutely. And I think it's exactly that. It's like a really great opportunity to now get creative, experiment with the stuff, and ideally overpass some of the competition. And I think that's also where the companies need to start looking at it of like, okay, if we wait too long, probably someone else really gets the competitive advantage edge out of it. Right. And that's why I think it should be something that the whole organization embraces. I'm actually quite happy that I think, where I'm at at the moment. It is an initiative that we're looking at from an executive level where we're like, okay, how can we enable all the different teams to experiment, to get access to the different tools, to license them and so on, so that we can look at how it can help our processes? And I think this is actually like a really good entry point into making it a little bit more practical.
Alex:
What are the things that you experimented with so far that were actually successful?
Ofir:
Yes, I think in my team we had a product we wanted to build out and I was experimenting with a lot of tools and we started using tools across the product management lifecycle. So the biggest thing I found generative AI tools are best for is starting new tasks. There's nothing more frustrating than looking at a blank screen or a document or a ticket or a board and thinking, how can I start this? Or a design canvas or something that is kind of staring back at you. And that's a very time consuming task to start, bootstrapping, everything. So we found it useful for doing a lot of the things for the first time in areas where we were not very proficient or they were new to us. For example, we went into a new market with a new product for developers and we used tools like ChatGPT and bing to do a competitive analysis and SWOT analysis of our competition and really identify where our strengths are and actually get a very detailed kind of a product vision board going on there. And we moved on to user personas. For example, we used a tool called user-persona dev to generate persona of the audience and actually get really deep into how a persona would look like, what's their goals, and what their pains are, and started actually looking at it actually generates images of people.
So it really helps you kind of connect in an emotional level to your persona. And also we started talking to our personas. I started having conversation with an imaginary persona and taking their point of view about their pains around the product areas that we were focusing on. Other examples were something that we often do writing product requirement documents and stories. So a lot of the times if you look at it, a story has a template that pretty much start off start and looks the same, but there is really some things are specific to that story and some things are just kind of more general. And there's definitely a way to generate user stories and product requirements documents with tools. For example, we experiment with writemyprd.com. It's a cool tool in natural language, it converses and you explain what you need from the product and then generate your outline.
User stories. To me, they blew my mind in how the tool, if you prompt it correctly, can actually deliver pretty good user story to build on top of. It won't be perfect, but it will definitely get you quickly up and running. And especially a lot of the times I hear from product managers that the engineering team is kind of pushing for more stories and it's very a race to put the requirements and designs in place just to get them started and even the initial ones, which are the hardest. The initial design you can ask any designer. The initial frame is the hardest and then it's copying and pasting most of the things and then iterating on top because the first button the first thing. These things we found very useful. I actually experimented with an auto designer tool called Uisard where you can prompt and get a full design out of just prompt and that includes clickable, mockups and even drawing on a napkin something and then uploading a drawing and then making it full working figma design from that tool.
So that's very cool and I kind of imagined where we're going with that. Another big thing that we saw very helpful was to write content inside the product. It was a coding challenge product. So we actually generated code sections and snippets without a lot of engineering. And then we gave the engineers only to fine tune and refine instead of doing the whole thing from scratch. And that's think about we want to do it in seven languages or ten programming languages, you have to find ten different developers to do the same piece of content, but the tool translates it automatically to all the different languages and that saves tons of time and resources. Another really cool thing was naming. I don't know if you've ever tried to name a product or a feature totally.
It's always a hard thing to do because everyone suggests a name and then emotionally connects to that name. For example, and we kind of said, hey, let's go with just kind of prompting for names and took ten names. And ran a survey in the team and actually went with a really good name that the chat actually offered just.
Alex:
As I came out of the exercise. A really good other addition to that is to then use the AI to validate those names. Because I think, as you say right, you can talk to the ten people. But with our recent branding exercise that we did for one of our sub-products, we had a couple of names and there were mixed feelings and we were like, okay, let's try with Jetgpt and see how could this name be perceived by a specific audience X, a specific audience Y or a specific audience Z? And we actually got quite a few interesting results back that really helped us understand and validate more the direction.
Christian:
But I just need to summarize what I'm listening to right now. I mean, there's like a project that you want to kick off and there's chat GPT, and there are certain other tools that you can tell to write your user stories, to write your product requirements down. You can define how the product has to look like. You combine everything you tell another AI to build a design out of this with clickable prototypes and then you can validate it against the machine. Again, that will give you very good feedback on whether this works or not. I mean, I'm just asking myself.
Christian:
What am I doing here? I need to stop podcasting. I need to build my own fucking product.
Ofir:
Definitely. So actually it's getting even crazier other than it doesn't stop in the product itself. It's also the go to market and the post launch where you can actually get the user interview scripts kind of started, drafted. You can get the copy for the product, for the marketing page. You can get the marketing videos. I experimented with a tool called did and I said, hey, I want a synthetic AI avatar of a developer talking about the product. And I said like t shirt geeky, 30 year old kind of. And then I gave it text and voice and it generates a fully functional human like marketing video, which is getting better and better and so on.
And really this is just the beginning. I think there's going to be a really big shift in how we approach product development after we experiment with these tools and after they really improve. Because some things like text is really advanced right now, but the video and audio and multimodal kind of evolution is yet to really commercialize and get in the hands of creators. This is really just the textual part where you really need to be a prompt engineer to do it. But it's going to a place where autonomy meets machine power. I don't know if you've heard of the auto GPT kind of autonomous agents, but these are agents that write their own prompts. They chain prompts and they get a task, a high level task, and they start to go deeper and deeper and deeper and write code and open web pages and start doing kind of actions in the real world. And that's really going to revolutionize how we work.
We won't have to be as technical and detailed as we are today in order to get a proper outcome. We'll just be iterating quickly over outputs from the machine and say, no, make it green. Okay, that looks better. Make it pink. We don't have to really imagine in the beginning how it's going to look like. And that will streamline the process of product development very much, I think.
Alex:
Yeah, I think another thing that I'm asking myself also when I hear all these things, I think we talked a lot about actually experimenting with AI and using it to your advantage and so on. How much should someone actually invest in not using it only as a productivity tool, but actually going into AI tooling and building AI tools for other consumers?
Ofir:
So I look at it, there's two aspects when I do a transformation AI transformation program for a customer or for my own team. I look at it from two angles. One is the productivity really achieving more with less, adopting tools and the other one is creating new value. And that is something that we need to ask ourselves as well. How can we integrate AI capabilities and put them in the hands of our users in order to empower them, to create new things inside our products and use the product in a way that is much more intuitive and really not only make our team more productive, but make our users more productive in the way they use our products and getting the value very quickly. Because every major tool will have an AI copilot assistant and that will become user expectation very quickly. It will become the expectation of the user. As product people, we need to think about that and understand it's now a competitive advantage, but in the near future it would just become a basic feature or a basic expectation out of essentially every digital product.
Christian:
Well, it's super crazy what is going on right now. And I think the more I talk about and also listening to you, the more I'm realizing how big this AI stuff really is for the whole product management world especially and product design as well. And if I look at the agenda that we discussed at the beginning, we talked about the AI revolution, we talked about tooling, but there was one more thing that you mentioned that is so important for you to also share with the audience, which is the whole thing about the AI mindset. And also to wrap up our whole conversation. I'm just wondering what can I do as a product manager or product designer to not only make use of the tools to stay on top of things, what can I do to really develop and dive into this whole AI work and develop a mindset over time. That is nothing that I have to do, but I want to do because it's so inspiring to me. Or beneficial.
Ofir:
So great question. I think the question I ask myself all the time is how do I not lose myself in the hype, right? So new tools coming every day and new use cases and they change and evolve and there's no way to really keep up with this. So really what I think is the AI mindset is basically understanding that AI is a powerful partner. It's a copilot, it's a brainstorming buddy, it's a force multipliers. And before we start a new task, we need to ask ourselves, can AI help me with this? Can integrating AI create new value? Or simply put, is there an AI for that? There's actually a website called there's Anaifordat.com where you can search AI tools and there's future tools IO and there's a lot of kind of tool databases that you can just go in and check what's the latest and greatest instead of kind of keeping context all the time in your head. It's just like Google searching. You're just searching for a tool, experimenting with kind of one tool a week just to. Get a challenge crushed with AI, see how it works for you and essentially see where it serves you best and where it's not really yet there.
And what I recommend practically is to first of all, check if the tools they're already using have or are going to release new AI features and experiment. Instead of writing my story, I'll just try to draft it with a tool. Or instead of writing my own user interview, I'll try to draft it and see if this is something that benefits me or not. And I will, of course, share with the audience all a few good resources that I use to keep up to speed and check whether these are evolved. One really good place is like it's called Product prompts, where it's very focused on product management tools and prompting and engineering. So I think the key is really to understand that we need to start experimenting quickly and learn. Because learn by doing is the only way you can master a practical tool. You cannot read about it in the news and then become proficient at it.
And that's the key.
Alex:
Yeah, absolutely. I think there's just like one final thing where I'm also curious to see how you look at it. But I think at the stage where we are now in AI, it's still important that there's the human factor controlling in the background. Hence also people from within a role like doing it. Because I think if someone never read a good user story or never or doesn't know the concept, the risk is high that whatever you get back from the eye is not the right thing. Today I've been playing around with synthetic users, which is like, I can do my six people user interviews and it gives me a lot of results. But still, knowing my audience personally, that allows me to triangulate the results, also shows me, okay, there is actually some things that are kind of off because the AI can simply also not know everything by me just giving them one single line of input and prompts. Right? So I think it's really good help.
It's a really good way to get started or to automate some mundane tasks. But at the same time, you cannot fully 100% trust it to do everything on its own and take things granted that you get back from it.
Ofir:
Definitely. So, yeah, definitely. We are in a driver's seat. We have to remember that especially now, it's an experimental phase. Nobody's actually releasing something that is not beta. Really. Everything is still they call it a continuous beta. Right.
They're always releasing early and putting it in the hands of users. And we need to be very responsible privacy of our users and of our own workplace. We have to be very mindful of the output and if we integrate it within the product, we need guardrails. So there's a lot of things we need to do in order to make sure that there's a person driving and AI is just copiloting and assisting, but not definitely not being cutting the human from the loop. I would not send out emails, automatic replies to my customers, or generate content and post it online without having an editor who understands because otherwise it would be a big mess. So definitely I'm not saying we back to the point. I was saying AI will not replace you, but a person using AI might. So be making sure that we are not staying behind because things are moving faster than we can imagine and there's no time to snooze on this one.
Experiment early, learn and make sure that we are in the right position and in the right time. I think the biggest takeaway I can.
Christian:
Give up amazing talking about no time for snoozing. I think it's not only for us, but also for our audience. So you can either now listen to this podcast or other podcasts, or you can just sit down and get your shit together. So press the follow button, continue listening to the product bakery and also work on your AI topics. No Opiate, it was a great conversation, really inspiring, and I'm really curious to see how much change we will just see over the next couple of months and years and I'm really looking forward to also stay in touch. And last but not least, I just want to say all the links will be linked in the podcast description as well as your LinkedIn profile. So if you want to stay up to date on what is changing in the AI world or you want to get in touch with ofYer, feel free to connect with him on LinkedIn as well.
Alex:
And of here, we got a lot of predictions from you, so I will also look forward to pick this conversation up again in the future. Look back at okay, how close did we actually get to your predictions and where do we stand in a couple of months from now?
Ofir:
Perfect. Thank you very much for having me, Christian and Alex. And thanks everyone for listening. And don't be a stranger, connect and check out my resources that I will be sharing with you. So thank you very much.
Christian:
All right, have a good one. Bye bye.
Alex:
Bye. Thank you.
