273: What AI Can (and Can’t) Do for Non-Technical Founders

Founders love the idea of skipping engineers and letting AI code their dream app.

Sadly, that dream falls apart fast.

AI can write code, but it also hallucinates, breaks in production, and leaves you with messes you can’t fix if you're not a coder.

What looks like a shortcut turns into a costly detour.

In this episode, Sophia Matveeva shows you where AI really helps founders—and where it doesn’t.

You’ll learn how to use AI to design prototypes, validate features, and impress investors without blowing thousands on guesswork. And you’ll hear why scaling, security, and serious launches still require pros.

In this episode, you will hear:

• Why AI won’t replace developers — and what it actually does well for founders
• How to turn an idea into a prototype that impresses users and investors
• The hidden risks of “vibe coding” every non-technical founder should know
• When to call in engineers, and how to avoid wasting money before you do

Resources from this Episode

Free AI Mini-Workshop for Non-Technical Founders

Learn how to go from idea to a tested product using AI — in under 30 minutes.

Get free access here: techfornontechies.co/aiclass

Follow and Review:

We’d love for you to follow us if you haven’t yet. Click that purple '+' in the top right corner of your Apple Podcasts app. We’d love it even more if you could drop a review or 5-star rating over on Apple Podcasts. Simply select “Ratings and Reviews” and “Write a Review” then a quick line with your favorite part of the episode. It only takes a second and it helps spread the word about the podcast.

Listen to Tech for Non-Techies on:

• Apple

• Spotify

• YouTube

• Audible

• Pandora

TRANSCRIPT

0:00:00 

If you're a non-technical founder, you might be tempted to think it's the age of AI and I can use AI to build my whole app myself. No developers, no hiring, no bills and this sounds perfect. Except that's not how it really works. Today, I'm going to show you what AI tools genuinely do well for non-technical founders, where they fail and what smart founders should do instead. 

0:00:37 

Hello and welcome to the Tech for Non-Techies podcast. I'm your host, Sophia Matveeva. If you're a non-technical founder building a tech product or adding AI to your business, you're in the right place. Each week, you'll get practical strategies, step-by-step playbooks and real-world case studies to help you launch and scale a tech business without learning to code. And this is not another startup show full of jargon, venture capital theater or tech bro bravado. Here we focus on building useful products that make money without hype and without code. I've written for the Harvard Business Review and lectured at Oxford, london Business School and Chicago Booth, so you are in safe hands. I've also helped hundreds of founders, both from concept to scalable product, and now it's your turn, so let's dive in. 

0:01:24 

Hello smart people, how are you today? In this episode, I want to help you make sense of the AI hype, specifically as a non-technical founder, because you've probably seen news headlines that AI will replace developers and that you don't need any engineers to build a tech product. And as a non-technical founder, this could give you false hope and take you down a wrong path, which could cost you actually a lot more money, because it's really tempting to think I could just build the whole thing myself using a chatbot without paying developers. And the thing is, there is some truth to how much AI has changed the game for product innovation, but it's not 100%. It's not 100% of that hype and that's what we're going to cover today. So by the end of this episode, you'll know when to use AI and when to call in the pros and why. 

So let's start with some definitions. You might have heard the term vibe coding. It's basically when you use AI like ChatGPT, claude, or there's the so called Ghostwriter by Replit to generate code from natural language prompts. So basically, you're literally just writing into the chatbot and you get code. So you tell it make me a website where users can upload cat photos and vote on the cutest one. I mean, I would be on that website all day and then AI would give you some code. So you paste it in, test it, tweak your prompt and try again. And this sounds great. 

But beware, and the thing is, this amazing promise, which I just told you to beware of, is backed by really well-paid PR departments from the world's richest companies. So, according to an article in the Financial Times called the Perils of Vibe Coding, microsoft now says that 30% of its code is AI written and Google claims the same. So what I am finding is that the way experienced engineers use AI is different to how non-technical founders use AI, and I'm assuming you're not an experienced engineer, and this is where I think a lot of the confusion and false promises come from. So that Financial Times article that I mentioned talks about a developer called Simon Willison, and he's well known for experimenting with vibe coding, ie basically asking AI to write code from his prompt, and he's actually created some cool stuff from code that he got from AI. So he created a software that summarizes blog comments and he even built a tool to alert him when a whale swam past his house on the Pacific coast. It's quite fun. So for Simon, he's an experienced developer. For somebody like him, these tools are a fun way to experiment because he's got the technical skills to fix the messy bits that AI makes himself. But listen to this In the article, simon said that he wouldn't use AI-generated code for projects that he planned to ship out unless he had reviewed each line. 

I mean you, if you've already been using ChatGPT or Gemini or Claude or whatever, you already know that chatbots, they hallucinate, they lie to us, right? So if you don't know how to spot the lie, if you don't know how to review code, then you shouldn't be using that code. It's the same with, like you know, drafting an email. So I use ChatGPT quite a lot, but I use it for subjects where either you know they're really really easy just like writing a very simple admin email or I'm using it for subjects that are areas where I'm an expert in. So if the AI takes me down the wrong path or it gives me some false information, I can correct it. So these are some dangers. 

But now let's look at how non-technical founders can and should use these AI tools. So last week I gave you a case study of one of my students, and here's a quick summary. So one of my alumni, noor. She runs two offline businesses a flower shop and a chocolate shop and she had an idea for a tech product to help small businesses take orders through Instagram. She had no technology background whatsoever and in the Tech for Non-Technical Founders course she learned how to use AI to design a test version of her product, so literally how to get the idea that she had in her head into something visual that can then be built into a real coded product. So this test product is called a prototype, and when she took this prototype to her users, she tweaked it. She ended up with a much better test product than she did, than she had right at the beginning, and this tested product helped her raise money from accelerators, which she then used to hire professional developers, and her product is about to come out. So she went from idea to launch in less than a year without blowing tens of thousands of dollars on guesswork, and this really is the sweet spot for AI and no code for non-technical founders. 

It's really useful at the very early stages, because at the very early stages, you need to figure out do people want your product? Or, for example, even if you know that they really want your product. You need to find out what do they want your product to look like, what features do you need to include in the first and the second versions? And also you know, even if you're sure of what features to include. If you give people a design, a visual version of your product, you need to make sure that they understand how to use those features right, because you know nobody's going to read instructions for an app. So these tools are very useful for validating concept and for getting early traction Because, for example, you could use AI to create a new test product. 

Show it to your target customers. If people are saying, oh my God, this is amazing, yes, I would love to use this, then you put them on your early adopter list. And so, basically, when you go to an accelerator or an investor, or when you're thinking of investing your own money into this, you're thinking, okay, I know what this product is going to look like, so I know what roadmap to give to engineers so I can get a proper budget from them, and also there's a whole list of customers that can't wait to use it. I mean, this is like this would have literally been a dream scenario for me a few years ago, when I first started getting into tech and these tools just weren't available. And, by the way, investors today they wouldn't take a founder seriously if they didn't see some sort of test product. So the days when you could just raise money with a pitch presentation they're disappearing because of these AI tools which you can use at the early stages. 

But let us be clear AI tools do not replace professionals. Think of AI tools like Excel. So spreadsheets did not replace accountants. You know, I still have an accountant for my business. I still have to pay them, but Excel changed the finance industry by making analysis easier and faster and more accessible. So no code and AI tools are doing the same thing for innovators today. They don't remove the need for skilled professionals, but they let you move much further before you bring the pros in, which is amazing, because this really de-risks your ventures, because you don't need to basically get money from somewhere either out of your own account or out of other people's to test a concept and to test features. 

And, by the way, if you want to get started with AI tools as a non-technical founder, then I recommend the free workshop that I made on this very topic for you. So in this workshop, you will learn how to use a free AI tool to make a test product. The full workshop is less than half an hour, so if you watch the videos and follow the prompts, you can have a product in your hands in literally an hour, and the AI tool that I recommend you use is also free. So the workshop is free, the tool is free and you know, if you have an idea and you've been pondering about it, then basically stop pondering and take an hour to create a version that you can show to real users and also impress your friends. So you can get this workshop at techfornontechies.co/AIclass. That's techfornontechies.co/AIclass, or just go to the link in the show notes. 

There are some more dangers of AI that I want you to be aware of as the business leader. So in that same Financial Times article that I mentioned, it said that developers who were using AI tools originally thought that they were doing work faster, but actually, when their output was measured, they were 20% slower than the people who were not using AI tools, and I don't know if you've had this experience, but I've definitely had this. When you know, I'm feeling a little bit lazy, so instead of drafting an email myself, I go to ChatGPT to do it, and then it takes so much prompting and so much work and so much editing with ChatGPT that it has actually taken me longer to do the thing than, basically, if I had just done it myself. So it's not necessarily a time saver. Sometimes it is, but sometimes it isn't. Also, I mentioned that AI hallucinates, so it makes code look right. 

That, then, doesn't work, and what I would say to you is that, you know, if you're still not entirely convinced, just think would you let ChadGPT write a contract for you buying a house? Like, if you're buying a house, are you going to hire a lawyer to draft a contract or are you going to get Chad GPT to do it? So it's too risky, right? It's too risky for something so important. You know, if you're handing over a big chunk of cash, it's too risky to basically just rely on AI. It's the same thing with creating a coded product, and so just remember what Simon Willison said. He said that he wouldn't use AI-generated code for projects that he planned to ship out unless he had reviewed each line. And you know what? I don't know this guy, simon, myself, but I have discussed this topic with high quality engineers and CTOs and they say the same thing. They're not just getting AI generated code and shipping it off as the finished product, because AI speeds up experimentation, but it is not great for production level work. 

So what do you do with this information? So the first thing that I told you about is that using AI to make a test product is a really, really good use of your time. Also, that's not the only thing. So, at the end of the day, tools don't make decisions, and you know these are AI tools, they're great tools, but they're tools. So the people who make decisions and you know these are AI tools, they're great tools, but they're tools so the people who make decisions are leaders, people like you. So this is why you're listening to this, so they don't tell you when to experiment and when to call in the pros. That's your job as a founder. 

So number one is, I would say know when to experiment. So use AI and know code to prototype. That's to make a test product to test your assumptions. You know. Test your assumptions about your users, test your assumptions about your features and validate whether your idea is worth pursuing, but also validate you know whether the features, whether what you want it to look like is actually what it should look like. Because you know the boss is not you, the boss are the customers actually what it should look like. Because you know the boss is not you, the boss are the customers. And you know that early stage that's kind of the cheap and cheerful stage, because that's when you can use AI to get feedback fast without wasting money. And, by the way, even if you already have an existing app or you already have an existing platform but you have a new feature, idea, like something really new that you want to add to it, then this is where creating a test version with AI can be super helpful, both on the time front and on the financial front. 

Number two as a business leader, you need to know when to bring in the professionals, because when you're starting to scale, when you're starting to onboard real users, and especially when you're starting to handle sensitive data, that's when you need engineers, that's when you need designers and that's when you need security experts. Basically, kind of think about it this way when things get serious, you need serious professionals. And number three as a leader, you need to know how to have the fluency to lead. So basically, that means that you don't need to be the one writing the code, or you know writing the code AI prompt, but you do need enough to understand how to manage a product team. You need to know who is on the product team. You need to know who does what on a product team. 

You need to understand their jargon, because they all have lots of jargon, because if you don't, you're not going to be a great leader and, in the worst case scenario, you could be led astray by people who basically just want to take your money. There are some of those out there that are also great developers who you know are going to act like your co-founders when they're just paid teams, who are going to act like your co-founders when they're just paid teams. But I would rather that you're saved than sorry, and so I would say that the real skill set for the age of AI, for a non-technical founder, is knowing what's possible with AI, knowing what's risky with AI and knowing basically how to push for clarity in this time of okay, there is lots and lots of potential for innovation, but there's also lots and lots of risk, and this is what I hope we got today in this episode and this you know. If you want to go further, then join TechPlan on Technical Founders, because that's exactly what we cover there. We cover not only how to use AI tools, but how to lead an innovation without getting duped by whatever is hyped right now, because do you remember a few years ago we were all supposed to be living in the metaverse and there were actually businesses that were trying to sell us virtual burgers. I mean, I hope you have never bought a virtual burger. Anyway, just a reminder. 

If you want to see exactly how to use AI to test your own idea, then check out the free AI mini workshop for non-technical founders, and if you follow the guide that I give you in the workshop, you'll have a test product in your hands ready in less than an hour, which is fabulous. So go to techfornontechies.co/AIclass or click the link in the show notes. Forward slash ai class or click the link in the show notes. And on that note, my dear smart listener, have a wonderful day and I shall be back in your delightful smart ears next week, ciao.