What was the specific moment or frustration that made you decide to build RepliStack?

When I was working as a freelancer, I received many repetitive questions from clients and potential clients. A lot of those emails were not difficult to answer, but they still took time and mental energy because I had to stop what I was doing, search for the right information, and write a clear reply. Most of the time, the answer already existed somewhere, in a pricing list, in a previous email, in a document, or in my own notes. In practice, replying often became a copy-paste exercise, especially for things like pricing details, service explanations, project steps, timelines, or standard answers to common questions. That frustration made me think there had to be a better way. I wanted a tool that could work with my own knowledge base and help me generate replies faster, without losing the personal and professional tone. Ideally, I wanted to tell the tool what extra details to include in a specific reply, and then let it generate a complete email based on the information I already had. Essentially, I wanted a faster and more reliable way to handle repetitive emails, so I could spend less time rewriting the same answers and more time focusing on actual client work.

What's one growth tactic that worked and one that completely flopped?

One uncomfortable truth I had to accept is that some people will prefer to use ChatGPT directly instead of paying for or adopting a specialized tool. At first, this can feel discouraging because ChatGPT is already widely known, flexible, and easy to access. Many potential users may think, “Why would I need another tool if I can just paste the email into ChatGPT and ask for a reply?” But this realization was useful because it forced me to think more clearly about what RepliStack needs to do better. A specialized product cannot simply be “ChatGPT in another interface.” It has to save more time, reduce friction, remember useful context, fit naturally into the user’s workflow, and produce consistent results with less effort from the user. For RepliStack, that means focusing on repetitive email workflows, saved context, faster replies, better structure, and a more practical experience for people who answer similar emails often. The product has to earn its place by being easier and more useful for this specific use case.

What do your customers complain about most, and how have you handled it?

So far, I have mainly worked with beta testers rather than paying customers, which has actually been very useful at this stage. Their feedback helped me see where the product was clear and where it still created friction. One of the most important pieces of feedback was that users wanted more control over the final reply. They did not want the tool to generate a generic answer only based on the email and saved context; they also wanted a simple way to add specific instructions for that particular response. Based on that, I added a new feature, a text box where users can write the exact things they want included in the reply. For example, they can mention a detail, a condition, a preferred tone, or an extra point that should be added to the message. This made the tool more flexible and more useful in real situations. I also simplified the login process in the browser extension because even small friction points can stop people from using a product regularly. In addition, I clarified the messaging on the website so visitors can understand faster what RepliStack does, who it is for, and why it is different from just using a general AI chatbot.

What's one thing about your market that most people get wrong?

I think many people misunderstand how much the quality of the result depends on the way the tool is used. A lot of people assume that if they open ChatGPT, paste an email, and ask for a reply, they will get more or less the same result as they would get from a specialized product like RepliStack. In some simple cases, that might be enough. But in real repetitive email workflows, the difference becomes clearer. If ChatGPT is used in a basic or mediocre way, without the right context, without saved information, without a clear structure, and without a workflow built around the actual problem, the results can be much weaker. The reply may sound generic, miss important details, or require several extra prompts before it becomes usable. At that point, the user still has to do a lot of manual work. I also noticed that many people do not realize how much effort goes into making a product feel simple. Stability, speed, privacy, login, browser extension behavior, formatting, and consistent output are not small details. They are the things that make the difference between a nice AI demo and a product that people can actually use in their daily workflow.

What's next for RepliStack in the next 6-12 months?

The next step is to finish the beta testing period for the browser extension and make sure the main workflow is stable enough for an official launch. At this stage, I do not want to rush only for the sake of launching. I want to make sure that the extension works properly, that the login flow is clear, that the generated replies are useful, and that the overall experience feels simple for the user. After the official launch, one of the main priorities will be to add support for more email providers. Right now, the extension works only with Gmail, which is a good starting point, but it also limits the number of people who can use it. Adding support for other providers would make the product more useful and more flexible. I also plan to continue improving the app based on real customer feedback. I do not want to add features just because they sound good. I want the next features to come from actual problems users have while replying to emails, so the product becomes more practical over time.

Before you started building, what was your relevant experience? Had you worked in this industry, built other products, or were you starting from scratch?

Before building this product, I had not built anything like this before. I did not come from a background where I had already launched multiple SaaS products or worked for years inside a software product company. RepliStack started more as a personal challenge and as a way to learn how to build a real product from zero, using the new technologies and AI-assisted development tools that are available today. The relevant experience I did have came more from the problem side than from the product-building side. As a freelancer, I had dealt directly with clients, answered many emails, explained services, sent pricing details, and repeated the same information many times. So I understood the frustration and the workflow quite well, even if I had not built a technical product before.

What's the biggest mistake you've made so far building this product, and what did you learn?

I think my biggest mistake was spending too much time in the build phase and making the MVP more complicated than it needed to be. I kept adding features, improving flows, and trying to solve too many things before getting enough real feedback from users. At the time, it felt like I was making the product better, but in reality, I was delaying the moment when I could learn from the market. I learned that it is better to launch a simpler version earlier, see how people actually use it, understand what they really need, and then improve from there.

If you could go back to day one, what would you do differently?

I would start differently. I would focus more on speed first, instead of prioritizing privacy from the beginning. I would also create the project files earlier, such as the handoff file, strategy file, and roadmap, so the project would be easier to manage from the start.