How to Break Into Product Management in 2025 (with Mariana Antaya, Product Manager at Microsoft)

Read The Transcript:

Hannah Clark: You know when you run into someone you haven't seen in a super long time and one of you is like, "so what's new with you?" and it's actually been so long since you've seen them last that you don't even know where to begin? That's how I felt talking to Mariana Antaya, who was a guest of ours back in February of 2024.

When Mariana was on the show back then, her career as a product manager had just begun, she just launched her TikTok account, and she was going on her third year as the founder of quantifAI, a crypto strategy optimization tool. But she came on with some super practical tips for aspiring PMs looking to break into the field which, to this day, is our most downloaded episode ever.

And right now, I can't even believe that was only a year ago. Since then, AI has gone mainstream, the tech job market looks completely different, and our respective platforms have grown significantly. So, we thought it was high time to invite Mariana back to the show to catch up, and more importantly, catch today's aspiring PMs up on what they need to know today in order to succeed as a PM in 2025. Let's jump in. 

Oh, by the way, we hold conversations like this every week, so if this sounds interesting to you, why not subscribe? Okay, now let's jump in. 

Welcome back to the Product Manager podcast. We have a very special episode today. We are joined today by Mariana Antaya, who was actually featured on our podcast a little over a year ago now. So for those who haven't been with us that long, I feel like such a mom, a proud mom telling this story. Mariana's not my daughter, but I still feel very proud because the way we met is because she reached out to us on LinkedIn. She commented on just a post about one of our episodes with an offhanded, "I would love to be on a podcast someday." 

And Becca, our producer and I were chatting and we're like, "let's have her on, let's see what she's about." And it ended up being our most popular episode of all time by quite a wide margin. So Mariana is a product manager at Microsoft now, and I'm talking about her like she's not right here. So thank you for joining us. 

Mariana Antaya: Of course. Absolutely. I'm super excited to reunite with you and be able to talk the latest and greatest in AI and product, and it's been a journey. We've both grown so much, so I'm super proud. 

Hannah Clark: Me too. Oh my gosh. 

Okay, so when we last spoke, we were just getting this show started. You were just getting your career off the ground. You started doing TikTok pretty recently. We had started recording and putting out episodes pretty recently. Both of us have experienced a very massive shift in our lives. So enough about us. Tell us about you. What's been going on in your career? What's been going on with your social media content? Tell me everything. 

Mariana Antaya: Yes, awesome. My career has been doing super well. I'm still at Microsoft and we've definitely shifted more to an AI product focus and strategy as most companies have as well, which has been really fascinating to see that shift really take place and grow within, the company culture.

And then on the content side, I am creating a lot of product content, but a lot of artificial intelligence content as well and really highlighting use cases of AI for businesses or use cases that my audience would really be fascinated in day to day. One of the most recent ones I did was predicting the race outcomes of a Formula One race, which is a hobby that I do myself.

And it was like, why not just post it out there and see if people would resonate? And it just seems like people are really resonating with all the coding projects and. The really cool and awesome use cases that AI has brought to light. 

Hannah Clark: Yeah, absolutely. And I am so excited to dig into this because it's like, we talked a lot to folks who are at the very top of their game, CPOs and VPs of product who are looking at implementing AI through an organization from a leadership lens. And now we're really talking to, you are really using it and in the weeds as a practitioner day in and day out. 

So let's talk a little bit about the shift in culture and AI's gonna, we'll have a whole bunch of stuff on AI, but just in general, when you think about how the culture of product was when you started in your career and now people coming in who are more junior than you, what's different now? 

Mariana Antaya: Definitely there's a big shift in the margin of where the technology has advanced. Number one. When I was just starting, I don't think ChatGPT was even a thing.

Now we use it every single day in our workflow. So that's one change is that the actual technology has advanced so much in such a short amount of time. And with that comes being able to incorporate and understand the use cases of how a big company can integrate that into our workflows and actually provide those use cases and solutions that customers will wanna use.

So there's a really big shift in culture from. More of a holistic approach to now. Also a little bit more of an experimental approach as well, and trying to iterate faster to better understand what use cases are customers willing to pay for and actually buy, or what use cases are actually going to land in this new space.

It is definitely more of an experimental and a faster moving pace from when I started. 

Hannah Clark: One of the things that we talk a lot about just as a culture and necessarily just on the show, is just layoffs and sort of volatility in the tech market. And I'm wondering from your perspective, do you see this as being like a time where people who are looking to get into the profession have more competition, or is it more just that people are looking for different skills?

Mariana Antaya: I think people are going to start looking for different and more niche type of skills for those use cases. If you're a domain expert, I think that's really gonna come to light in this era because. The specific prompts or the specific use cases that XY, Z target market is really gonna resonate with.

So if you can really dial in on a niche that you have domain expertise in, or you have hands-on experience with, from my perspective is what's gonna make you like a very successful PM in that market. 

Hannah Clark: That makes a lot of sense. Let's move into the AI stuff. I know that we're both like nine to talk about it. We'll chunk it out into a few different sections. Let's first just focus on your day-to-day. We're, we'll talk about building with AI in a moment. 

You mentioned that ChatGPT is part of your day-to-day workflow, and I'm sure that it's infiltrated a bunch of different functions. What are some of the different ways that you're using AI to get your day-to-day work done?

Mariana Antaya: The biggest use cases I use it as is as a market researcher for me and to do competitive analysis for me. So now I no longer, need to wait for a researcher to get that information for me or a big company that gets contracted to do all of the market and competitive analysis, or even myself, it takes me less time to go infiltrate the customer or the competition and kind of see what new features they have.

I can just prompt an LLM to construct a market research paper for me to tell me what the gaps my product is missing versus theirs, and also to understand how our product can be ahead of the curve as well. That's one of the, most practical use cases also for building product specs. AI is not going to take away our career as product managers just because they can write a product spec.

Even some of the LLMs, you still need that human interaction and component because we're the ones physically also talking to customers like all day and every day. So being able to have that human touch is important, but it gives me a really good baseline for the product specs that I'm writing, whether it be a new feature or multiple features.

So that's another really high touch use case that I love using AI for. 

Hannah Clark: Yeah. Okay. I think we're finding new ways every day. We've had some folks come on talking about using AI as a way to be more effective at parsing data from user interviews and different ways of kind of training it to respond as a user persona.

Yeah, it's really innovative, the different ways to support your workflow that just weren't possible before. 

Mariana Antaya: Even creating agents. So creating agents is super awesome. I was able to create an agent to basically build out a lot of the dashboards that I have for my features. So now I'm, I no longer need to spend hours building out dashboards to see all the user metrics, which is something that, as a product manager, we heavily rely also on the data to back up our hypotheses.

Yeah, building agents for. Things that you do really repetitively has also been a game changer. 

Hannah Clark: Yeah. We did a great episode with Tal Raviv a little while ago on how to build an AI copilot for product managers that like, took us through the process and like just hearing it from him was just like, you can really do this. It was, yeah it's so cool. 

And just moving into the skill sets, because obviously, once we're in the field, we're gonna be using it all the time. But how do we prepare if we're currently, let's say we're talking to aspiring PMs, I'm sure that there's more than a few listening who are wanting to know, what do I need to kinda be familiar with, or what should I get comfortable with before I start applying for jobs in which I'm going to be using AI all the time?

Mariana Antaya: I think one of the most relevant skills you can learn, especially as an aspiring PM, is prompt engineering and being able to experiment with the different LLMs. And what use cases are best for LLMs? For example, like when I'm writing code, I really love to use the Claude. I love using Claude Sonnet 3.7 is awesome.

For writing code versus I really prefer something like ChatGPT in order to do competitive analysis or road mapping, or even helping me structure my prompts for other LLMs. Actually, I will prompt that ChatGPT, hey, for this LLM, how should I prompt to get an optimal result for X, Y, Z? One of the best, I think, and most valuable skills you can learn as an aspiring PM is prompt engineering and exploring and even building your own product. 

There's this whole phenomenon of like vibe coding now as well, and that's like an amazing way if, even if you're a non-technical PM, to really put your product skills to the task because you can literally just prompt an LLM, create a product and try to build a go-to market strategy, try to sell your product.

Try to go to customers and ask them about their experience using their product. And that's real life product experience that you're getting that isn't really taught in schools. And that's really valuable experience that is gonna translate to the real world as well. So I think that's definitely one of the key points I wanna touch on there. 

Hannah Clark: Yeah, absolutely. Yeah, I'm very excited about this kind of new vibe, coding trend. I'm taking a workshop in it later this week. I'm very excited to dig into. So we'll get into prompt engineering shortly. 

I wanna talk a little bit about building AI products. So the other side, the more customer facing thing. You touched on vibe coding as a way to get started with building products for customers. But building AI products specifically is a whole other thing that's changed in the AI or in the product management landscape. You're currently doing that, so what is that like? How can you prepare for building AI products?

Mariana Antaya: I think a big part of it is understanding that everyone is on this journey together right now as well. We're all experimenting. We don't quite know yet at the moment, like what are the big use cases that people will really gravitate towards and what customers will actually are willing to pay for and at what margin they're willing to pay for that feature or that product at.

So one, we're all in this together, but two is really understanding more of the experimenting with models. I think there are so many different models at our disposal as a PM and collaborate really closely with engineering so that they make sure that accuracy is really important or latency as well.

We're seeing that now that there's more complexities within softwares that's going to. Increase the time that things appear on the screen, for example. And so really understanding the risk to reward ratio of how long is your user willing to wait there to get that additional benefit to them will be important.

And that's the role of the product manager to step in and really analyze as well. Keeping a really close touch with your customer will be really important there. And understanding their workflow, maybe a four or five second delay for them to see some extra AI generated summary doesn't work for their workflow, so being able to understand those challenges or really important.

Hannah Clark: Yeah, that's a good point. 

Let's move on to prompt engineering. You mentioned that this is like a key skill. I think this is not just a key skill for our PMs who are looking to, or folks who are aspiring to become PMs. This is like a new skill set for everybody and I don't think anyone can be too good at it.

So what's your like prompt engineering 101, like when you're thinking about building a quality prompt, what's like the main thing to keep in mind? 

Mariana Antaya: The more information I say and the more context that you can write in the prompt, the better. If you're very vague with your prompt, you're most likely going to get a very vague response in return.

So being able to really either show your reasoning step by step is a great way to prompt better. So just say, Hey, analyze X, Y, Z, and step one, step two, step three, or even formatting your prompts can help in that respect. There's also. Newer reasoning prompts that have recently come out. So understanding what type of model you want to use for your prompt will also be pretty important, but being able to even ask the prompt for additional details.

Don't be afraid to tell the LLM, like it did a bad job. Correct the LLM in that case in order. To get the prompt that you want. Or like I mentioned earlier, ask a different LLM, Hey, how am I able to get the most effective output of this LLM? And there are people who have written papers on that as well.

So definitely being able to explain and structure your prompt as well as give it as much context as you can possible all of the documents, add in images, add in the books, add in whatever you need into the brain so that it can understand where you're coming from and the context that you need. So I think that's one of the most efficient and effective ways that I've been able to get better at prompt engineering and understanding also that.

You have to tell the LLM. Tell ChatGPT, think of yourself as a data scientist or think of yourself as a senior product manager at x, Y, Z company. So giving it an identity is also a great way to structure your prompts that way it starts thinking more so in that vein.

Hannah Clark: Moving into iteration on stuff, 'cause you mentioned, correcting LLMs. Something that I've been getting in the habit of doing that's really been serving me well is closing a feedback loop with them, where if it gives me an output and I make modifications to the output before I use it, then I'll give it back here's what I've done with your last output.

Please remember this so that you can give me like a more efficient output or a more accurate output later. And that kind of tends to nudge it in the right direction, which I've found to be really useful and an easy way to do that. But what kinds of tips have you picked up or little like tricks up your sleeve for getting better and better outcomes every time. 

Mariana Antaya: I think I'm still on that journey and path as well, besides giving it more context, like the tips. I shared, I'm still learning. I'm not an expert in prompt engineering myself. I definitely just try to iterate on the response, or sometimes I will try to add in the same prompt into a different LLM and see what the response is.

Playgrounds are especially useful. For this because now there's a tool I use, or even in GitHub and NES has this as well, where you can compare the output of two different models side by side. So that's a really great way to better analyze which LLM is giving you a more accurate response or a response that's more tailored to what you want.

Hannah Clark: Oh, that's a really great tip. Yeah, I suppose that's how you can discern like what you know, whether Claude is the better fit for a specific task or ChatGPT. I was actually wondering about that, was there a reason, are there nuances to why you decided to use ChatGPT for more like analysis versus Claude for writing code? And did you notice like one thing particularly in those specific use cases versus the others? 

Mariana Antaya: Yeah, for sure. So for example, in the coding example, if I'm trying to code with ChatGPT, a lot of times if I get an error response or if I get an error after I run the code and then I enter the error into ChatGPT, sometimes it like won't fix the error.

It'll take multiple tries for the LLM to fix the air versus Claude, I have a much higher success rate. I just find that it writes a lot cleaner code and more concise code. So that's why I prefer Claude. But see, these are little nuances that unless you really try and experiment with LLMs ones you don't know.

Hannah Clark: It's interesting 'cause I'm finding the same thing. I'm almost finding that Claude is a more of a creative thinker if that's. I feel a little weird using creativity in the same vein as LLMs 'cause it's so like un comfy territory. But I do find that what I've used it for, it's more effective for using or for developing marketing collateral versus ChatGPT I've, yeah, I've agree.

I've found it to be a little bit more efficient or just give better outputs when it comes to research or using it for more kind of an analytical work. Why do you prefer ChatGPT for on that same vein, like why have you found that ChatGPT is better for more analytical tasks? 

Mariana Antaya: I find that the research it can do is very in depth as well, a very in depth two.

It is more accurate and the way the research is outputted is very customizable. I can tell it, Hey, I wanna chart or give me a paper. It will more so like reason through what I'm prompting it to do. So it's also a bit of personal preference for your use case and for what you are going to use it for and your role too.

There's definitely a personal preference in there makes it. 

Hannah Clark: Yeah, I definitely agree. We were actually talking about this internally and a lot of us were just fans of how Claude uses sepia tones. Like just something simple like that, just like this UX little tiny shift that's it's just easier on the ice, so it does make a difference. 

I won't keep you too long, so thank you so much for taking a break from your busy life and chatting with us and catching up. But for those who can't get enough and wanna chat with you some more or wanna follow what you're doing, where can people find you online?

Mariana Antaya: Absolutely. You can find me on LinkedIn, Mariana Antaya, or always on TikTok and Instagram @mar_antaya. I post a bunch of coding product, AI videos on there, so.

Hannah Clark: Cool. Alright. Thanks for being on here. 

Mariana Antaya: Always happy to chat and collaborate with you. If you wanna coffee chat, open to talking, just feel free to reach out to me on LinkedIn.

Hannah Clark: Thanks so much. Thank you. 

Thanks for listening in. For more great insights, how-to guides and tool reviews, subscribe to our newsletter at theproductmanager.com/subscribe. You can hear more conversations like this by subscribing to the Product Manager wherever you get your podcasts.