Ep 38 | Prompt Like a Product Pro in ChatGPT & Claude: The 3-Phase System you must know (w/ Bastian Epskamp)

Want clearer, more consistent AI outputs? In this episode, we break down three prompting skills that help founders, PMs, and UX leads go from idea → clickable MVP fast—using ChatGPT and Claude. Guest Bastian Epskamp shares his 3-phase system (prepare → implement → learn), how to ship prototypes stupid fast with V0/vibe-coding, and a case study on building Auction Buddy (AI for buyer-side property decisions).

What you’ll learn

The 3 skills: framing, constraint design, and iterative prompting (ChatGPT & Claude).
Prompt templates that work: turn vague asks into structured tasks for reliable outputs.
Speed to MVP: V0/vibe-coding, docs-as-leverage, tight feedback loops.
Real product examples: how “Auction Buddy” scores inspections, plans bids, and does live auction maths.
Team craft: pair human creativity with AI to reduce cycles and increase signal.

Who this helps: Founders, product managers, and UX leads shipping LLM apps, agentic workflows, internal knowledge assistants, and customer-facing AI features.

Timestamps:
00:00 Intro — Meet Bastian (BE Innovations) & the “product lens”
03:25 Why AI is a product superpower
05:40 Early builds: Dreamweaver sites, custom PCs, first “maker” wins
08:45 Research & prompting in practice (talk to AI like a teammate)
16:05 Teams, mistakes, and building with intent
19:30 Exit story → starting a strategy/innovation lab
21:00 Monetisation & prioritisation: pick the sharpest pain points
24:30 AI = remove repetition, keep judgment (why experience still matters)
29:10 Aussie ecosystem: capital cycles, resilience, and pivots
31:00 Upskilling across functions (seeing the whole system)
35:30 Prompt frameworks that work (structure, constraints, examples)
38:20 Rapid loop: Prepare → Execute → Learn
39:10 Property tools: “Auction Buddy” & decision support for buyers

👍 If this helped, like, subscribe, and share with a builder who needs faster MVPs.

#ChatGPT #Claude #PromptEngineering #AIProduct #MVP #LLM #Prototyping #VibeCoding #Startups #UX

Transcript

Speaker: 0:00

Today's guest, we have Bastian Boskamp, founder of V innovation. I always look at everything in How could you build a product but also monetize it? They are constantly trying to this new environment, and a lot There are two types of afraid of AI, and the other one use AI for everything. Identify your biggest pain something where AI. If you can use AI and train an model with a data set that's can basically help get knowledge business, but you still need assistant, so to speak, right? How do younger people probably prepare themselves for the market? A just understand things change If you're able to pivot, you Because every problem in life and in business, you need to pivot in order to stay successful. The second part of it is AI is Bringing these different views I split my prompts into three One is like the pre work where I like we ask AI a lot of hey, what did you understand? What questions would you need to Then I do the implementation prompt itself and now implement it. Make sure you follow XYZ rules. You have an incredible guess that intersection of part of your life now. Um, do you mind just like giving product, uh, expertise and how in this space over the last importance and the skill set you Sure. Um, I guess I always look at How could you build a product but also monetize it? Yep. And that has always been my view Right. Um, we can dive into this later, wrecks just because I wanted to building physical products. I have a software engineering background, and anything I look at is basic from that product lens. And there's there's a few industries that really interest me, but everything is more or less about how can new technologies help? How can you know new things make things in product space? Build more products have been early on in blockchain, worked in that space for a while before everyone actually knew what it was. Yeah. Uh, just because it's I think can do a lot. And with AI, it's the same. I've been. in there early on and I've gone gone through that journey of GPT. Why don't you follow my It's still kind of doing that Yeah. But you know, you learn along prompt, how do you use things? And with that, I realized that AI is an incredible tool for especially products and that kind of. Like early stage of like trying to find market traction, MVP working with different people in. An environment to express what product is is that link between. Your customer service, your tech You know, the C-suite, And you need to speak different languages to explain it in ways to different stakeholders and groups. Um, and that generally means you context and focus in that sense. But also you constantly try and write things down and explain them, but AI helps to speed those things up that are just repetitive tasks and also kind of help you to refine it in a way that it's more understandable. WIP coding even though the word like it is. It's really cool where you use tools to visualize what you're trying to do. MVP coding is so much more I built probably like a clickable prototype every two or three weeks. Not necessarily because I need to, but I'm also interested in it. And that is where I really comes perspective, from a creativity things up so that I can focus on to spend time. That's awesome. Make sure it's your, um. What's interesting there is that of being an engineer. So actually being on the tools Um, often you see a lot of product people transitioning from, I guess, the design or UX backgrounds and um, or just going straight into product in itself. Why did you transition from, I a product, and how have you Yeah, I think there are two I started playing computer games, you know, you Counter-Strike. Oh, yes. Whatever. We can have a huge conversation That's awesome. And back then, I mean, most people our age might understand it, but otherwise all of a sudden you have your blue screen. Something doesn't work, right. And you need it to build your How do you connect a graphics All that stuff. And it started making me want. Or it made me want to understand Yeah. Um, on top of this, I started a my neighbor trying to build put them into emails. We stopped this because who Right. This is this is like years and Um, yeah. So basically, I always try to And when it came to, what do I I was thinking of studying art. I'm not an artist at all, so I Or studying software software engineering, and that I've done only a few years. But what I realized is rather than just coding, not just coding, but coding, I wanted to be the one that's in charge and responsible for the good and the bad. Yeah. And we all know that product people usually blame when things go wrong or that don't get praised. I've never done that. I don't know what you're talking No, but this is this is going into product. Because why choosing between art Maybe there's still that Yeah. yeah, just wanting to be cool that elevates a product know, USP in the market. Um, and that evolves into more than just product management but like strategy. So I do a bit of consulting, um, for like small and medium from a different angle. And what's interesting also just to to finish this up is the market seems to come around from product. People need to be creative, need have a design background, to People that need to be And I while if you want me to code something right now without AI, it's probably going to break. And it's taken years. But I can understand code and I can talk to tech people equally as well as to, you know, the business side and understand this and then bring those two things together. That was the critical part that There is that the ability to happening from the development product people who have they've They're great at the strategy But when it comes to the development building, there's a huge disconnect. The why of what is going on is Um, I myself am in a similar boat while not being on the tools. I was a producer and product Like actually, like I helped. I was the guy who managed all So while I don't have the technical skill that you have, I do have a good understanding of how that core development process is working, what is going on, why we're using certain codes for certain reasons, and that does help with that intersection of how you deal with stakeholders, how your product becomes more robust when it goes into market, um, which is which is a great skill to have. So that's awesome. Yeah. And just to add here, I think that also helps me to kind of I'm not an expert, but I understand how AI takes how it thinks. Yeah. Because you talk to AI or I try to a computer. More technical. I structure my prompts. I structure everything from a than just talking like a human. And it makes a huge difference And that's the same when when you talk with engineers, if you say, I want a product so that people can record their email address, right. It's different. If you were to say, we want to build a email capture field, but we're going to have ten thousand users coming to this website every day. What does that mean for What does that mean for the What does that mean for privacy So you can frame these things a And that technical knowledge. Apply this to AI. Even though coding you don't I still believe it's beneficial to know those things because you can usually prompt more specific and get better results in general. Usually it's that when you start prompt, it's even going to the extent of what does good not look like. You know, when you're trying to insight, as well as all the dive into a lot of prompt I'm pretty excited to talk to Um, along your journey, you've different, uh, I guess, uh, Tell me about your, um, your stopping point when it came to the the the surfboard, uh, stuff. So, yeah, that's an interesting So I used to when I came to Australia eleven, twelve years ago. Eleven years ago, I started surfing and living in a shared house. All the surfboards were It's strange, as an Australian, Oh, I don't I'm fearful. I'm fearful of it. I just I don't want to go in the I just like, ah yeah, it's fun, but I'm not claiming to be able to. Yeah, yeah, of course I'm probably the kook that's on that surfboard annoying everyone else. But anyway, so I looked at this And they're like, oh, I want to be able to kind of like store my surfboards. But they're not just flying on the garage floor and like getting dinged and stuff like that. And I looked into basically products that were out there and There was one specific product from the US to have, like a surfboard rack to store your surfboards that was beautifully designed, but it cost like, I think eight thousand nine hundred Australian dollars, like we're talking seven, eight years ago. Yeah. And this is the economic side. If you take inflation into whatever twelve hundred now. Yeah. Um, there was another company, this better also, which. Yeah. Um, also, can I really do it Different things. I wanted to understand the being responsible for digital product strategy, you're able to with physical products. And in the end, I worked with an surfboard racks, built, um, and So that's a bit more sustainable Yep. Not talking about shipping and Yeah. Um, but yeah, had that idea of like, okay, I'm taking the idea from Ikea. Flat packs. Ship them straight to customers. Um, and it worked okay. I mean, I spent my own money, learned so much on the way. Like, before importing them, I get asked, so where's your fumigation certificate? Like, what's your fumigation Applying this back to, like, like to test because all this make better strategic decisions, product and then ultimately the Yeah, right. And did you what was the lessons know, it was such a niche thing these bamboo boards, which opinion, particularly when what type of boards they're. But you, you I mean, you ended Yes. Um, from a what did you learn that digital environment that into, I guess, how you position Yeah. Interesting question. I think one of the things I've bottom line, because if I would you know, I wouldn't have worked Like, there's no salaries for startup people and expect the unexpected. Just just to give you an idea. Um, I get all these boards on, Where do I store them? I thought I'm going to store shared house, but all of a percent of the garage. Uh, you know, like, I haven't And taking this, this experience of, like, think about the unexpected. Be prepared, um, think outside And also have a good network of A friend of mine, he has a to help me understand how the I had no idea. Right. Like try and import like a container full of stuff into Australia. Um, yeah. And taking that into the digital world, it it made me understand working with different stakeholders. Again, you need to have that important for them, but also If I think this works and this everyone else likes it. And even if everyone else likes right thing and then turn it This is where I is cool. I'm building something right now don't have a designer, I don't I have no one else but myself prototype and be able to test it Yeah, because digital is In AI helps me to be quite fast, level in terms of how the Um, and therefore I'm able to focus on the important things like you've gone from, I guess, like selling these racks, selling these surfboards, but like going through then the health industry you worked in medical director during the Covid pandemic, which is incredible. And they were responsible for, I telehealth support for How did you find that sort of, I time versus the work you were Because I guess there was a lot There was a lot of people who fear that was in the market and a solution for these people in a How did you find that Um, incredibly beneficial for my Yeah. However, it was obviously We had, I think medical director had like forty seven percent Percent market share of patient management systems for GPS in Australia. Roughly. Don't quote me on this, but I'm amount of GPS would rely on it. And when you get told we need to know if people otherwise die physically to a doctor, infect Um, it was a lot of pressure. Um, but we had an incredible So it's not just me, obviously, and everyone kind of knew the direction. And this was more than just a It was this helps people. And back then nobody knew what's I mean, I think we started in something just before we went And yeah, that, that that was an I it's going to sound weird, but positive things from it. As in, if you work in a team, you have a purpose as in your There's so much more energy to useful, because of course you positive purpose, right? As well, of course. Yes. Purpose in the sense of it's Yes, it benefits society. And of course there's always the Yeah. Of course. Um, yeah. That was why people profit. Yeah. And you know, like before the old discussion between working from home or going back to the office. Now, if you think about it, other than a day here and there. Yeah. And all of a sudden, people have They have their kids with them. Life got so chaotic, yet everyone was pulling in the right direction. And like, we built something together, which just means again, if if you believe in your product, if you believe in what you can build, and if you have a good network and people that work with you and you work with, you can achieve some amazing things that that that's basically my main takeaway from me. What, um, what has that given starting your own business and because you got be innovations? Yeah. Um, you're in consultancy where consultancy to people, Um, as a leading strategy point. How have you found that has guided you into where you are now? Um, I think it's not just the work at Medicaldirector, but I've worked in a lot of different industries. I've worked in blockchain early on, um, for probably three years. Um, I worked in health tech. Obviously. I've worked in, in direct I worked in big data. Um, and all these things kind of like helped me to understand Different aspects of product building strategy. What to focus on. Also, and this ties back probably to medical director make hard decisions on what to focus on versus what's not that important, right? Ignore all the shiny things. If you don't do this thing like you know, are you going to survive? Is the product you go going to Yeah. And I worked for a lot of startups and scale ups, some successful. That sold for quite a nice Um, I was just employed. Yeah. Um, and others renegotiate. Renegotiate? Yeah. Like, that's that's the thing, Like, some were very successful. Some weren't. I was, um, in one startup. It was the first hire ever. And other companies were had But I was the first product hire a very technical team and culture and strategy. All this helped me to understand zero to, you know, one hundred. So to speak. And yeah, that combined with I always like to build stuff and experiment. That's what led me through it. And funny enough, just to start my own strategy business approached by a company to do Product transition, if you will. And yeah, that just led one They were very happy with my I'm like, this could lead to into what I like to do. You elevating a product, you looking at a specific problem or an overall problem and bringing it in this experience from different industries, software engineering background, physical products, digital products, all these different things, and then basically have all these ingredients and make a cake, so to speak. Yeah, it's it's perfect. I like it. Tell tell me about BD like what is it that you guys For people and for businesses, I Sure, it's a range of different One is more looking at strategy Either build products in you early on as a startup and prioritize and focus on the right things? Or how can you strategically How does the market look right What's your potential user base? How can you monetize it? Just the overall strategy of And the other thing is looking at things like product transition. You have too many products. You have, um, you know, a vendor you work with that's not the best to work with and you don't know how to basically get things back on track, as well as building MVPs with AI to bring this back. Yeah. Um, so that when you have ideas, current team with just someone from the outside that set of mind or like a fresh Um, yeah, that's that's pretty I love it. I guess it's it's great. I mean, it's it's what the And I think we're seeing a lot over the place, as well as large through these big product The impact driven by AI, even even driven by market needs and change in economics, um, around the globe, they are constantly trying to figure out how to adapt into this new environment that is rapidly growing around them. And there are a lot of them are suffering and the opportunity is there. It's ripe. Yeah. But, you know, then the question faster to adapt with what's us supporting them. So I guess. Yeah. How do you see? What do you see as the pains that they're suffering for the most? And how do they overcome those? Um, generally speaking, I think the biggest pain is that if you are an established business, it doesn't matter if it's small, medium, big, you, everyone in the environment or most people in your environment has a specific thinking, specific mindset. Yeah. Right. Because if you work with you don't have the time to look ten thousand feet outside view. Yeah. Um, you also might be very industry specific in terms of your thinking. Um, but someone from the outside things that have worked for in different environments. That's one thing. The other thing is, I think there are two types of businesses. one that are very afraid of AI, and the other one is that really, really wants to use AI for everything. Mhm. I strongly believe there's probably something in between where identify your biggest pain points and see if that's something where AI can help you solving them. But because the biggest pain a lot of risk. Do it in a call it R&D fashion as in just an easy example, let's say your customer service team is overwhelmed with repetitive questions. Yep. And yes, this is an easy Their solutions are it's common. It's quite a common example Especially with if you're not keeping up with what's going on, there is always constant complaints coming in from your customers. Fix this I want this, I need Like how does the customer Um, get their own knowledge. If you can use AI and train an model with a data set that's can basically help this customer from every system in your engineers document, what you say to salespeople. Yeah, there's I think it's Q it's kind of like an internal And I've just gone through this yesterday where I think AI, at the end of the day, is a tool to take away repetitive tasks or help you look at data in a way you would usually not be able to, as in connecting A, B, and C together. And that's something that's incredibly important for businesses. But you also need to be mindful do it right, what happens? Right? Like, what's the impact of it? And that's why I'm also not a everyone and AI will do the job. Yes. And that's something that we're Um, there are people, businesses, organizations, particularly large corporates who are getting rid of product teams, UX teams, design teams and attempting to replace them with AI with gigantic, you know, technologies. Um, is this something obviously, like, what's your view? And like, why shouldn't they be How is human in the loop such an a product person or UX person, Yeah. Well, first of all, I don't Yeah. Um, also, everything that I'm saying right now might not be relevant, might be not relevant in like six months or twelve months because things change so rapidly. It is. But if we take today A kind of I think the change is that you your organization and more in the junior level. Um, but you still need experts so to speak, right? Like, I always see AI tools as out of uni. Yeah, they are very promising Um, they have an amazing skill They will have very brilliant produce, but because they don't unfortunately, sometimes you and basically need to question And if you look at this. If you look at this from like a lot because AI can produce keep an eye on it. And businesses need to keep this want to reduce their workforce. There's also something that if you use AI tools every day and just auction buddy, I'm building it. Um, I use for example, v0 for my If you use V0 for a while, for different MVP's, for different clickable prototypes, for different industries, you kind of start to see a pattern of it's always a square box, it's always this. And that's where like a human perspective could change that in a different way. And so, long story short, I market, they will have a more the need for less. Look at my product. I don't have a designer, no building everything for a usable if there's some traction. Right? Yeah. Um, that's easily like three, would have needed to hire and more senior environment. And I personally think even up understand how it works. And if you understand this, position your business in a way successful and you're going to If you don't, you might not use use it at all. And that's where you fall So you've got like, um, things the founders of Canva, founders know, we're not letting go of impacts of AI. We're letting them go because of increased demand and needs. We've been able to automate a lot of flows that aren't to do with AI, just specifically with the technologies generally that we're implementing. Um, but on the flip side, you have, you know, founders of businesses like, um, uh, Shopify who are coming out making reports going, we're not going to hire people unless you can prove to me that AI can't do that job. Um, so how when you have, you are coming in fresh to the supposed to enter a market where an economics point of view, and AI can do your job. Like, how do they make particularly in product? The statement I think the CEO of Shopify, that was three or six months ago. It was. Yes. This is like in AI terms, this is like a decade in anomaly, right? So so it would be interesting to Um, I also think reading between in everyone in the organization, So you like ahead of the curve and we have the best talent possible. Yes, there are certain tasks replace a human. And this is the second part of Um, in terms of how do younger themselves for the market? I think a just understand things And I've seen this in blockchain You wake up and something Regulatory changes, whatever and Yeah that is crazy sometimes and But at the same time it also you think differently. Because every problem in life and in business, you need to pivot in order to stay successful. So you learn something out of it And the second part of it is, I think, experiment, if, you know, fingers crossed, maybe somebody gets successful and I need to and can will be able to hire some people. I would hire people that use AI. And when I say you say, show me what you've built in your private time. It doesn't have to be pretty. It doesn't have to work. But show me that you have interest because I, at this point is still a junior assistant. Still needs to understand how it There are people saying, oh, you don't need to understand how prompting works anymore because there are tools that can prompt for you. That is correct. But I want people still to understand how to prompt, because that structure makes a world of difference in the results. And if you know this, you can then, you know, like basically upskill yourself. So if everyone built as much as you can, it's the idea of, you know, as technologists ourselves, particularly product people, UX, that all that strategic technology and digital. Our role is to stay ahead of understand how the product needs And if we're not staying ahead of the curve and what's happening in AI, we're making ourselves irrelevant. Essentially, half of the work exactly if I wouldn't use AI. And also one thing is, I think more, as in human design. But that doesn't mean that you're not a product person, because AI helps you to kind of like be a bit more of a product person. I'm a product, but doesn't mean certain degree, because AI helps So it's it's blurring the line. And it's interesting that I've seen you talk about is that know, the need for structured also the intersection of Um, how do you see that working, where there's a lot of question Um, I for one, one of the biggest areas of creativity that I see the most when it comes to running any of these type of processes is, is the different roles that people can play in the development. So whether you're a finance or you're a CEO or you know, started that difference. And diversity brings so much any kind of process and project. And now, with AI on top, that's completely changing, but also kind of leveling the field with that mentality and that creative element. How are you seeing those two I think creativity lies in Creativity is something that AI The creativity that AI does is And when I say you, I mean like what it got trained on is in the large language models are trained on. Right. Yeah. Um, creativity for me is also being able to see things from different angles and connecting the dots. Which brings me back to a Might see things more from a This would be beautiful and The I don't know, the like, but this doesn't make us would be like, but this takes one only one month. So you bring in these different creativity I think. AI. Perfect. I love that helps. Like I would love companies sometimes to build your call it like your rapid strike squad as in one designer, one product, one dev. Take the vision or the mission free with AI and build something Right? And the other thing is just to creative as what we feed it and how we think and how you Of course. Right. Will it be relevant that things perfect in five years time? Maybe not because nobody cares Maybe I don't know, but right And that's where Having different people in in your circle of friends, different demographics, different mindset, different everything as well as in your professional life makes a difference. Which brings us back to Think outside of the box. Yeah. Awesome. That was that was really well I liked that a lot too. That just the creative element an important piece for me. Whenever I run any of these product roles, whenever I run a project, whenever I help customers develop startups, develop something. Diversity is always a crucial was well summarized. Thanks. Um, obviously in practice, um, I piece that you mentioned. As we as we dive into that, there's a lot of, I guess, naivety and how we structure the prompts that or how people who are inexperienced structure prompts. Um, and you've come out, we had around the importance of that style and how you structure that So we're touching on that How do you see prompting being such an important piece of the output that you need it to produce? I got to be bold and say prompting is ninety percent of what quality of result you will get. Yeah. Um, like I said in the ChatGPT the way I would use build me a website to do email Nowadays what I've learned is Format in a structured way. Yeah, but also I split my prompts into three different components. One is like the pre work where I where I explain things, these kind of things. Um, and then ask AI to come back with what it understood, what, what the potential outcome would be if this would be used to implement a new feature functionality. Whatever. What does that mean to risk? Ask me questions. So it's a two way street for me, Like, we ask AI a lot of hey, what did you understand? What questions would you need to designer, a software engineer? So give it a role. The result I'm getting from this understand how AI thinks, based terms of information, and then it before I implement something implementation prompt itself and Make sure you follow XYZ rules. Make sure you don't touch this. Remember last time we had you Remember it's for this customer And yes you can outsource this Like this is a mobile first So seventy percent of users will How does that impact your coding Whatever. Once this is done, the third and last problem is what have we not AI? But what have we learned from As in, should I have prompted What has I learned in it? And that will be then documented in a rules file like for auction. But I think my rules file is now Um, and I have multiple files, rules for specific projects for So I don't need to type it every itself takes the most time, and it's going to be. Right. Yeah. And then that's where things go So prompt structure prepare. Execute. Learn. Nice. Touched on a couple of times your your auction platform that you're building. Tell me about this that sure. Because it sounds really What you've uh what you've been hopefully by the time this be launched and ready to be Yeah. Auction. If anyone wants to hear first. Yes. Um, basically, again, the always looking at things from a I build and how could you day, how could it help users? And the idea is the property market in Australia is very much Um, except from buyers agents, you don't really have anyone that supports the buyer upfront with informed information and that product. The idea is you, as a potential buyer, go to different properties that you are interested in. Inspect them, record information for you specifically of your interest, um, specifically for this property. And once you have gone to like you all of a sudden realize whatever, this needs fifty it's One hundred thousand that doesn't need the repairs. Um, so you get more and more tailored to you. You then set a bidding strategy You equip with this property similar to XYZ and your tailored information. You know, um, you should bid up live auction day. What does one hundred thousand for your down payment, for your yield and everything else? And it's basically supporting information, because you go auctions and see information. And then obviously, yes, there's Yeah. Um, to help you make the right informed decisions because property in Australia is very emotion driven. A lot of people overspend. A lot of people Regret, probably I mean, you know, there's that's available and therefore overbidding because they've been they've been struggling to get And then suddenly then it's out Hundreds of thousands more than exactly what they're willing to stretch themselves. And the buyout. I mean, the buyer can either go an Excel sheet or somewhere, or a lot of money. This is an alternative where it's free, potentially some premium features. Again, I take a picture of a room and say, what would it cost to renovate this in a country house style, which medium to cheap furniture? And it can basically calculate And all of a sudden, you see, I want to renovate. It cost me an extra one hundred This is cheaper than getting a And then obviously, from a could then connect you to shops selling shares, whatever. Yeah. Um, yeah. And I've done the business case. I've done everything. I think there's there's I've I had some people already Well, we'll make sure to include Mate, that sounds really awesome because it's definitely something that people are needing out there in this current market. Bastion. It has been an absolute pleasure Any final words that you want to Um, no. Just anything you want Com.au give me some feedback. There will be a whitelist for five hundred people to have it Other than that, just thanks for This is amazing. I've been a pleasure. Thanks. Thank you buddy. Cheers, everyone. Thank you very much. Like subscribe and we'll see you