Piyush Jain, Founder and CEO of Simpalm and co-founder of Ducknowl, is on a mission to solve real-world challenges by combining technology and entrepreneurship. With over 15 years of experience building custom software solutions, Piyush helps businesses turn complex ideas into practical applications by blending technical depth, business acumen, and a strong problem-solving mindset.

We explore Piyush’s AI Ideation Framework—Validate idea, Proof of concept, Design, Competitor analysis, and Feature selection—a practical approach to building software in the post-AI era. Piyush explains how AI can help teams better understand user personas, validate product assumptions, and rapidly prototype ideas, while human expertise remains essential in design, architecture, and production-grade development. He also shares how prompt engineering, peer-reviewed prompting, and a right-shoring delivery model can help businesses build smarter, faster, and more cost-effectively.

Good day, dear listeners. Steve Preda here with the Management Blueprint, and my guest today is Piyush Jain, the Founder and CEO of Simpalm, a custom software development company, and the co-founder of Ducknowl, a candidate screening and assessment application business for high-volume recruiting. Piyush, welcome to the show. 

Thank you, Steve. Thanks for inviting me. 

Well, I’m very curious about the stuff that you have to share with us, and I’d like to ask first about your personal purpose. What is your “why,” and how are you manifesting it in your business? 

Yeah, so that’s a very interesting question. And I think for every entrepreneur or tech founder, really, that’s the motivation—why you want to do certain things. So for me, if I look at it, my personal “why” is: why are we not solving challenges? Or why are we not solving them the right way? Why are we not transforming our lives? I grew up in India and then came to the US, so I’ve seen many different parts of the world—from Asia to North America. I see people face different challenges, but then we are not focusing on solving those problems. A lot of it I see is there’s a lot of challenges in the world because I believe there are not enough entrepreneurs. Because entrepreneurs are the ones who really take risks, combine everything, and create solutions. That was like me, right?

That’s what I learned growing up, that I think I can do that, right?

I can combine the technical knowledge and the business acumen and create solutions that people like, solve their challenges. Growing up, like I'm more on the technical side.
Share on X

I like it. So solving challenges and being an entrepreneur, and kind of combining the two—being the technical expert and the entrepreneur in one. Now, one of the things that we always talk about on this podcast is frameworks. And you have developed a really good one for AI ideation, which I think is something that everyone needs to do these days or use these days, and it helps you create business apps and other business applications. Can you share with me how that framework works, and what are the steps in it? 

Sure, yeah, definitely. So just to give you a brief background, we’ve been building software for the last 15 years. Some companies have used different frameworks, whether it’s Agile or Waterfall in SDLC, in building the software, right? There are different methodology that companies have used, and they’ve been good, successful—they’ve played their role. But now,

with the advent of AI, things have changed. We had to figure out, in our organization, how to use AI, and that’s how this framework was built. My team helped me building this framework as well.
Share on X

How did you do it? 

So really, we built this framework—very interesting. A lot of the steps are similar, but then a lot of things are different.
Share on X

For example, maybe we are building a healthcare application for an anesthesiologist. I don’t know much about that. I know, I mean, because I have been through some medical surgery and all that, but I can’t fully understand their user persona or their requirements with respect to the application we’re building. But now, with AI, I can actually ask different AI models, “Hey, we are building this app for anesthesiologists. What are their pain points? How would they see it?” So all that deeper mindset and psychology we can get using AI. 

You are validating the idea by interrogating AI applications. 

What users are going to like and all that. So I will always use this term earlier. In software engineering, now we have this pre-AI and post-AI, right? If you read history, we talk about before Christ and after Christ, right? Yeah. So it’s a similar thing now. Yeah, exactly. Or before Covid, after Covid. Before AI, after we did all the user research and everything and created a requirements document, we would usually do design, create like a visual design of the software.

But now, with the AI framework, we don't do that. That's not the next step. What we do instead is create a quick prototype using AI platforms.
Share on X

So there are a lot of AI platforms—like Lovable, Claude. Now ChatGPT launched Codex for coding, and Replit. Depending on what kind of application you’re building—for example, maybe if you’re building a web-based application—then I recommend using Lovable or Replit. They’re very good at creating that. Whatever software you want to build, whatever user personas that you’re addressing, you can feed into that and it’ll create like a prototype application.

Okay. 

So what that does is actually, then this prototype, clients can just take it to their customers or internal users and get feedback. A picture is better than a thousand words. Organizations discussing an idea is very different from when they actually see something. Then everybody starts chipping in—“Oh yeah, I see this in the prototype, but I don’t want this,” or “I want to move things around,” or “This is what I want.” Basically, building a prototype on AI platforms is much faster than building wireframes and design prototypes like we used to do earlier. So that has changed.

So you’re 3D printing your software, right? 

Yes, exactly. There you go. Well, that’s a very good way you put it together. Yeah. So, yeah, exactly. You’re just 3D printing the software, right? So you can see it, visualize it, and then once you go through that, it creates a lot of better ideas about the software in faster time. So once you have that, then you go into UI/UX design. So in that also, there are two steps. One is wireframing. Wireframing is like creating the flow in black and white. It’s like creating a skeleton of your software. It does not have the color, the font, or the branding, but you just create all the different user journeys, the screens, the flow, and the fields that will be there on the screen. So we have integrated AI into that step as well. Earlier, it used to be created by a designer or a business analyst. Now we are using software like Uizard or UX Pilot, where we define what we want—what kind of user journey, flows, and screens—and it creates that. It spins out those wireframes in minutes. So really that has reduced now. The time it used to take to create wire frames is faster now. 

So you’re designing the wireframes with AI? 

Yes, but it’s just the wireframe part of it, and it’s still guided by our expert VA or designer—someone who knows how to really visualize things and has done a lot of wireframes and sketches. So they know what to tell the AI. Prompting is very important. It’s very important that you know how to prompt—what to ask for—so that you can get variations and differentiation in the wireframes. You don’t want a standard AI-created wireframe. Everybody can recognize AI-generated images now, right? If I show you one, you’d say, “Oh yeah, it’s AI-generated.” I know that, right? Yeah. So again, we keep the human intelligence. We’re not asking AI to create the full software end-to-end. It never works—it’ll never work. It just doesn’t. I know that’s a strong statement, but I’m saying that based on experience and an understanding of human behavior and psychology.

So AI agents will not be able to code software, in your opinion? 

No, they can do the coding, but they cannot build the whole software end-to-end—a production-deployed software.

Because these software are being used by humans. You have to have human intelligence to understand and define what you need and how it works.
Share on X

So just to recap—you validate the idea by interrogating Claude and ChatGPT, asking about the needs of that customer, the psychology of the customer—that’s step number one. Step number two is 3D printing the software with Lovable or Replit—so proof of concept. And then you design the wireframes. And then what’s next after you design the wireframes? What’s the next step? 

So that’s a good thing. That’s it. Now I’m going to talk about the human element—some people listening to this podcast will be surprised. Now it comes to visual design, right? So you’ve created the skeleton, and now you have to add the skin, the tone, the color, the emotion to the design, to the workflow. Now, we have tried AI, but it doesn’t work. It’s very monotonous. So we use an experienced visual designer, a UX designer, for that step—to give it emotion. When you use AI—I wish I could show you some examples—it creates very similar kinds of designs for apps and software. So what we did is we gave it three different apps with very different objectives and everything, and the designs it came up with were very similar—blocks, buttons—very monotonous. So there’s no differentiation. And design is the main thing that becomes the differentiator, right? 

Yeah. 

So that’s what we learned from our experience. And I say that very categorically in all of my talks—that visual design, final UX, has to be human, not AI.
Share on X

Yeah. It doesn’t have emotions. 

Well, some people will argue with you and say, “No, it can understand if you’re sad or unhappy.” But my response to that is—it’s because we’ve programmed it that way. But things change based on situation, context, ethnicity, culture, fear—how people express nervousness, fear, and all that—it’s very different. So there was this AI video interviewing company five or six years ago. They were sued by the Department of Justice because they were trying to detect emotions of people like anxious, nervous, when the interview was happening. 

It turned out their model was trained only on one race—they didn’t account for other races or ethnicities. So their model failed, and they were sued by Department of Justice for that. So yeah, emotions is something—maybe they have unlimited dimensions, we don’t know. So it’s hard to program that. So basically: ideation, prototype, wireframe, and then final visual design—that’s the discovery and design framework. Now, when it comes to development framework, this is where AI has been a game changer—the coding part.

But again, you have to be very careful about how you use AI in your coding pattern with your coding team. It depends on the application, it depends on the tech stack, right? Every platform has its own strengths and weaknesses. For example, if you want to build a web-based application in the React JS framework, then Lovable is great. That’s very good—very efficient and cost-effective. Then Claude is there. Claude has been really good in software engineering. I would say it has been built and designed mostly for coding, right? Anthropic—their idea, their starting point—was coding, how to make coding and software engineering better. 

So they’ve been a front runner in the race. ChatGPT is trying to catch up using Codex, and Copilot is great. Copilot is mostly used by enterprises who are on the Microsoft stack. They use Copilot a lot for coding in .NET and enterprise-level applications. They’re used to co-pilot. It’s because they feel comfortable with Microsoft security policies and all that. That’s fine. But in general, we see Claude to be at the top—from our perspective.

We’ve also built a framework for software coding. In software development, there’s a popular process called peer review. So when you create source code, you get it reviewed by your peer—your colleague.
Share on X

Is this what happens on GitHub? 

Yeah, yes. So basically anywhere—any source code repository—you can do that. So your team members can help you make your code better and more efficient. 

Yeah, I understand.

But now we have a step called prompt peer review. When you’re using prompts to build software, those prompts get reviewed by team members. Because if your prompts are not very specific or good enough all the way through the SDLC, you can run into a lot of challenges trying to fix the code. Because now you have a situation where you have code that you have not written fully, and when you ask AI to change something in the code, sometimes it ends up changing a lot of things that you don’t want it to change.

Yeah. 

That’s what we’ve seen, and that’s why we evolved. Before we build any software, we create maybe a 10-, 20-, 30-page prompt document, where we go through each screen and function and write it out. It’s very sophisticated—it has evolved really well. But the thing is, it takes a few days to do that within the team, because we know if we do it right, the next step is faster and more accurate. So really, the prompt document—think of it more like an architecture document. Earlier, we used to create a solution architecture document, defining all the tools, the design, everything. 

But now it’s more like an AI-driven solution architecture document with prompts, which get reviewed by team members. So we do that, and then we run that, and we get the code and everything. So I have a CTO club—I run a CTO Club in Maryland—and I was talking to CTOs. They’re all using this, but some of them are so advanced that they actually define the test cases in the beginning. They define, “Okay, this is what I want, this is the function I want, and these are the test cases I want it to pass.” That’s even more advanced. If you can do that, you can have very efficient code. 

Yeah, I love it.  So is that the end? You have your test cases, you design the prompt, you peer-review the prompt, and you already had the prototype, so now you’re coding the software—what’s the last step? 

Yeah. Then there’s an integration as well. So AI doesn’t do the integration so well. You can do the front-end coding, you can do the back-end coding, you can probably create the APIs. APIs require a lot more human intervention. But once you have that, then you have to connect it, right? You have to connect the front end with the backend. A lot of that is still done by the programmer. It’s hard to rely on AI for doing that. And again, it depends on the application. Maybe if it’s a smaller application, maybe you can have AI do that. But if it’s a bigger application—we mostly build bigger applications—then integration, then final QA and testing, and deployment. 

So all that is there. But in each of these steps, you can use some sort of AI tool to speed up the process. But the key is you still have to have your architecture, the process. You have to know the steps more. You have to be a good, experienced developer to use AI efficiently if you want to build a production-ready application. You can build a prototype. Anybody can build a prototype on Replit or Lovable, but it’s not going to be production-ready that you can give to your customer and charge them money. So that’s the differentiator. 

Yeah, I understand. So Piyush, I’d like to switch gears here. I understand the AI ideation framework—that’s great. We talked about the technical part of it, the curiosity, the technical challenges. Let’s talk about the entrepreneurship part, which is also part of your profile. So what drives the growth of your business? What would you say drives it? 

For us, there are multiple factors that drive the growth of our business. The first is, again, our problem-solving attitude. Any client that comes to us we communicate in that model
Share on X

Some people call it hybrid shoring. I call it right shoring. The reason I call it right shoring is because in this model, you have the right people at the right shore, so you get the most value. Here, you have people who understand the culture, the product, the context—because products are used by people in a certain culture. And if you are not in that culture, if you haven’t experienced it, it’s always harder to design the right software solution. I was one of the first people to start that model here in the DMV area for mid-size and smaller companies. This model existed before, but mostly for large enterprise companies. They have used that. But I started to offer that 16 years ago to smaller companies. Either companies were just going offshore, or they were doing onshore, right? I introduced this hybrid—or right-shoring—model, and it has been well received by our customers. So that’s it. 

So what is one thing that you’re trying to figure out in your business right now? 

Right now, what I’m trying to figure out in my business is scaling. I mean, we have built solutions for many different industries. We have built solutions for different clients in fintech, healthcare, education, nonprofit, startups, IoT, construction. But now what we are trying to figure out is how do we create some off-the-shelf solutions for different industries? Because one challenge we see is that, from the client’s perspective, getting custom software built takes time and money. But in certain use cases, we can have off-the-shelf, industry-specific solutions, and then customize those based on the client’s needs. 

So that’s what we are trying to figure out—across different industries, what those solutions can be—so we can scale and also make it easier. And these are more like AI-driven, off-the-shelf solutions that are customizable. So think of it like Salesforce—its core is off-the-shelf, but then you can customize the front end and a lot of other things. Not exactly like Salesforce, but more like industry-specific solutions for different use cases—nonprofit, construction, right? With those, overall, we can build solutions faster. 

That’s fascinating. So how has the offshoring—or right shoring, as you call it—model evolved over the past 10 years? Is it different now than it was 10 or 20 years ago? 

Yeah, I think that’s a great question. It has evolved and changed. Earlier—maybe 10, 12 years ago—when we were talking about hybrid shoring, we were mostly talking about the US and Asia. But now we have different players. We have the nearshore model, which has become quite popular as well—like South America. We have team members in nearshore locations as well, in South America, because we want to leverage different time zones, resources, and culture. And we’ve seen very positive results. Then you have Eastern Europe. We have competition from countries like Ukraine, Belarus, Romania, Poland. I think it’s the part of the globalized world, right? It’s like energy flowing in different spaces—it’s not limited to one place, which is great. That’s one way it has evolved. 

I also know some companies working in Kenya—there are developers there. Some companies are setting up in East Africa, West Africa. So different places are playing roles now. That’s one thing I see. And now, with the help of AI, what’s going to happen is it will play two roles. One— in many situations, with AI, you can do more things onshore. That’s one aspect of it. And second—with AI, someone sitting offshore who knows how to use AI can become very competitive as well. We don’t have enough data yet to fully see how this will evolve, but maybe in a year or so, we’ll see how it plays out. 

But I also find that with these simultaneous translation tools—like Apple, I think an iPhone can now translate in all languages. Essentially, another barrier falls that if the language and knowledge of your offshore contractor is not perfect, they can understand things much more clearly because of simultaneous translation. Even on Zoom, you can now flip a switch and they can read what’s being said in their own language during a conversation. So that’s amazing, I think. 

Yeah. That’s amazing. That’s amazing. They can understand more about the culture and mindset. So that’s something have to see. Again, I think it depends on the use case, the application, the problem we’re solving. But in some cases, it might be great news for onshore—we can keep more dollars here. But keeping dollars here with AI also means a lot of that spend is going to AI, right? So that’s one thing—we have to be very careful. Yesterday, in our tech breakfast, our presentation was about how to optimize your AI tokens. There are some companies spending $150,000 per year per employee on tokens.

Wow. 

That’s like the salary of one employee. 

Yeah. 

A mid-level developer—$150K—they’re spending that much. And then they’re trying to figure out how to optimize it. And on top of that, they have cloud costs, right? AWS, Azure—those costs are still there—and then you add AI. So it’s a lot of money. You really have to be very smart about understanding and optimizing it. That’s why the prompting is so important, right? It’s not just about getting the right software—it’s also about getting the cost down. 

Yeah. Again, you need expert people who can prompt well, because it’s about being able to communicate well. Prompting is about communication—it’s about clarity, brevity, security, all that stuff. So, Piyush, we’re coming close to the end of the recording. If someone would like to learn more about the applications you develop, how you’re using AI, and how you can help their business develop technology, where can they find you? What’s the best way to get in touch with you?

Sure, there are many ways people can reach out to me. They can go to my website, www.simpalm.com—we have a contact form there. They can submit the form, or they can reach out to me via email directly at [email protected]. They can also connect with me on LinkedIn. I’m on LinkedIn—message me there if somebody needs anything. I always like discussing problems and what the solutions can be. If anybody reaches out to me, I’m always very quick to respond. 

That’s awesome. So Piyush Jain, the CEO of Simpalm—and we didn’t even talk about your other business, Ducknowl—thank you for coming, and thank you for sharing your insights and your framework on how to build an ideation framework for AI. So thanks for sharing that. And if you’re listening and you enjoyed this conversation, then stay tuned, because every week we have another entrepreneur sharing their insights and frameworks with you. So make sure you follow us on YouTube, subscribe, and give us a review on Apple Podcasts. So thanks for coming.

Thank you, Steve. It was a pleasure talking to you.