Interview: How does AI support the creation of Custom Software?

When developing custom software, the key decisions are made at a very early stage, often before a single line of code is written. It is at this point that not only the system’s functionality is defined, but above all its real business value. Until recently, work at this stage involved analyzing documentation prepared by business analysts or, at best, relying on static and time consuming visualizations based on that documentation. Today, Artificial Intelligence enables significant changes in this process.

We spoke with Katarzyna Lewicka, Head of Business Analysis & Quality Assurance, about how combining interactive prototypes with the conscious use of AI is transforming the way projects are run in their early stages, accelerating decision making, and enabling the delivery of better tailored solutions.

Why is the discovery phase so important in the custom software development process?

„In the case of custom built software, we are not creating a ready made, universal solution. We create applications tailored to a specific business, its processes, and its users. This means that the decisions made at the beginning of the project have a huge impact on everything that happens later.

The discovery phase makes it possible to properly understand the project context, not only the scope of the planned functionalities, but above all the problem that needs to be solved and how it fits into real business processes. Without this, it is easy to create a solution that is technically correct but does not deliver the expected business value.”

What role do interactive prototypes play in this process?

„In custom software projects, documentation alone often proves insufficient because it describes something that is difficult to clearly imagine. This is especially true for clients who may not have a technical background or experience working with prototypes, which makes it harder to properly interpret assumptions and evaluate proposed solutions.

An interactive prototype takes the discussion to a completely different level. Instead of describing the system, we can demonstrate how we understand it. Users and stakeholders can actively interact with the prototype in a way similar to working with the final application, navigate its structure, recreate key processes, and better understand the logic of the solution by relating it to their everyday tasks. As a result, discussions become more specific and decisions more informed.”

What risks does this approach help reduce?

„The biggest risk in custom software projects is not the technology itself, but whether we have properly understood the client’s requirements and, as a result, whether the solution we are building is actually the right one. Very often, the problem is not that something was implemented incorrectly, but that misunderstandings about the client’s expectations occurred during the analysis stage. As a result, even if the implementation itself was technically correct, the process or functionality did not meet the client’s needs.

Prototypes make it possible to verify this very early, before development begins. They help us identify discrepancies in product understanding, process mismatches, or solutions that may be unintuitive for end users. This significantly reduces the need for costly changes in later stages of the project.”

Does this approach affect the project cost?

„It depends… Using prototypes does not suddenly make a project dozens of percent cheaper. What changes, however, is the efficiency of the entire software development process. By investing more attention at the beginning, we can minimize the risk of errors in later stages. As a result, the project is better aligned with the needs of the client and end users, which reduces the need for costly changes or fixes later on.”

And how does this translate into delivery time?

„Since key decisions are made earlier and based on a clickable prototype, something concrete and tangible, projects can be delivered more efficiently. The team does not need to revisit core assumptions or reinterpret requirements during development and can instead focus on delivering functionality rather than constantly correcting direction.”

Where does AI come into this process?

„We treat Artificial Intelligence as a supporting tool that helps us work more efficiently during the analysis and design phase. AI makes it easier to prepare alternative solution variants and accelerates certain aspects of the process, while decisions still remain on the analyst’s side.

What is crucial is that we use Artificial Intelligence consciously. We do not use it to lower project costs at the expense of quality, but to deliver higher quality without requiring the client to increase their budget. An approach focused on cutting costs through excessive reliance on AI is, of course, possible, but in practice it leads to client dissatisfaction, a greater number of revisions, and, in the long term, an increase in the overall project cost.”

So AI does not replace the team?

„Definitely not. In Custom Software projects, the greatest value lies in understanding the business context and being able to translate it into a working system. This requires experience, dialogue with the client, and responsible project decisions.

AI helps us streamline this process, but it does not replace competence, experience, or human intuition. With AI support, we are able to work faster and deliver better thought out solutions.”

Does this approach change the way you collaborate with clients?

„Definitely yes. The process becomes more iterative and collaborative. Instead of spending a long time working with abstract requirements, we quickly move to concrete proposals that can be viewed, tested, and discussed.

The client has a greater influence on the project direction from the very beginning, and we can respond more quickly to their needs and feedback.”

How do clients respond to working with prototypes?

„Very positively, because it gives them a greater sense of control over the project. They can see the system before it is built, test it, and make sure it meets their expectations.

This is especially important in projects where the investment is significant and the level of alignment with the organization’s needs is critical.”

How would you summarize your approach?

„When developing custom systems, the goal is to ensure the solution is the right one. That is why we focus on making good decisions as early as possible, reducing risk, and delivering a product that genuinely supports the business.
Interactive prototypes and the conscious use of AI are tools that help us achieve this goal, not by taking shortcuts, but by planning the process more effectively.”