Vasyl Soloshchuk, CEO and Founder of INSART – Interview Series

Vasyl Soloshchuk, CEO and Founder of INSART, is a fintech entrepreneur with more than 25 years of experience across software development, engineering leadership, and venture investing. He started his career as a developer and steadily moved into leadership roles, ultimately scaling INSART into an international player focused on financial technology. Throughout his career, he has worked closely with founders and executive teams to shape product strategy and execution, while also contributing to the broader tech ecosystem through initiatives like the Kharkiv IT Cluster and participation in venture syndicates supporting emerging startups.

INSART positions itself as a fintech accelerator, combining engineering expertise with strategic support to help companies build and scale financial products at every stage of growth. The company supports startups and enterprises from early ideation and MVP development through product-market fit, go-to-market, and large-scale deployment. With a focus on areas such as digital banking, payments, lending, and financial infrastructure, INSART acts as both a technical partner and a growth platform, helping fintech companies accelerate development timelines and navigate the complexities of scaling in regulated markets.

You began your career as a software developer; what early signals convinced you that financial services would become one of the most technology-driven industries?

I started as a freelance software developer when I was still in university. I quickly got a team together and we began taking on projects through freelance websites. Software development and finance both really interested me then. I was doing software development. Learning about finance at the same time. It was a way to combine my skills and interests.

What shaped me early on was the combination — building software, studying finance, and thinking about financial freedom and entrepreneurship.I became quite obsessed with the intersection of tech and finance because I strongly believe that people can be free only when they are financially free.

It was also a very crucial moment in my life.

Back in 2008, I was a PhD intern at IBM Watson Research Center in New York. That was exactly during the financial crisis. I was working on data analytics projects while also closely following developments in the financial markets. I was in New York, visiting the Stock Exchange and talking to people at IBM who worked with hedge funds and investment firms. There was a real sense of confusion and frustration — people did not understand how to move forward. That was the moment when I realized that the financial industry, as it existed, was not stable. It needed to be redefined. And for me, it became clear that technology would be the driver of that change.

I had a chance to stay at IBM. But I chose to do something else, I decided to build INSART. My goal was to help create technology. I wanted to make the financial system stronger and more open. Now that I think about it there were some signs that made me make that choice.

First, financial services are fundamentally information systems. Money is data, transactions are state changes. Once you understand that, it becomes obvious that software is not a support function — it is the core of the business. Companies like Stripe are a good example of this, where APIs essentially became the product.

Second, regulation was increasing in complexity. Every new requirement adds layers of logic and reporting. This cannot scale with people alone. It requires systems. You can see this clearly in companies like Plaid, which built infrastructure to handle that complexity.

And third, margin pressure forces automation. Financial institutions compete on speed, cost, and risk. Technology is the only lever that improves all three. Over time, working with fintech companies, we saw that the winners were not those with better ideas, but those with better architecture and execution. And that is still the mission we follow at INSART today — to advance fintech businesses that redefine finance and expand digital freedom.

Where is AI already delivering measurable value today?

The gap between AI as a narrative and AI as a production system is still very large.

Where we see real value is in areas with strong data and fast feedback loops.

Fraud detection is the clearest example. Companies like Feedzai and Stripe Radar use machine learning to detect anomalies in real time, and the impact is directly measurable in reduced fraud losses.

Operational automation is another area. For example, companies like UiPath are helping financial institutions automate document-heavy processes, reducing costs and speeding up execution.

Personalization is also working well. Wealthfront and similar platforms use AI to tailor investment strategies and improve user engagement.

One area that is emerging very fast is generative business intelligence. Many companies already have data, but leadership cannot access insights fast enough. We are building systems where executives can ask questions like “Why did revenue drop this week?” and get immediate answers across systems.

That is where AI becomes very practical — it improves decision speed, not just processes.

In underwriting, fraud detection, and compliance — what is actually working?

Fraud detection is already working at scale. It has clear signals and continuous feedback. Companies like Featurespace are strong examples of this.

Underwriting is partially working. It performs well in certain segments, especially when using alternative data. For example, Upstart uses AI for lending decisions, but still operates within regulatory constraints.

Regulatory compliance is the most complex area. Much of what’s touted as AI is still rules-based automation. True AI-powered compliance is more complex, requiring contextual understanding rather than mere patterns. Companies like ComplyAdvantage are heading in this direction, but it’s still evolving.

There is also a quieter but important shift happening in internal reporting and decision-making. Many institutions still rely on static dashboards. AI is starting to transform this into dynamic systems that enable leadership to understand what is happening in real time and act faster.

How much of the current AI excitement is hype versus real capability?

AI is both real and over-marketed at the same time.

Part of it delivers actual production value.  The rest is still narrative-driven, especially in fundraising and positioning.

The difference is simple. If AI improves a measurable KPI, it is real. If it mainly improves a pitch deck, it is hype.

For example, adding AI labels to a dashboard isn’t transformational. But if AI helps a CFO instantly understand revenue drivers and take action, that’s a real opportunity.

Companies like OpenAI and Anthropic are building powerful core models, but the real challenge is how to apply them to business systems.

That’s why we focus on decision intelligence and generative business analytics. This is one of the few areas where value is immediate and adoption is natural.

Why do so many AI pilots fail to reach production?

Most failures are connected to systems not to models. The fundamental problems are fragmented data, lack of production infrastructure, and unclear accountability for results.

Additionally, AI is developed in isolation, so teams create models and these models are not integrated into workflows or reporting systems. A good example of it is banks that are experimenting with models similar to those offered by companies like DataRobot but are unable to implement them.

If the data that comes out of this process does not reach the C-level decision-makers who can use it, then the whole project comes to a stop. AI models only work when they are part of how we make decisions, not when they are something we are trying out on the side.

How do institutions overcome data infrastructure challenges?

It is not about  AI, but data architecture. Well-performing institutions unify their data, create real-time data processing pipelines, and set consistent metrics. Companies like Snowflake and Databricks are really important in the area because they help ensure data infrastructure can handle large volumes of data. And this is the foundation to build strong decision-making systems and reporting. 

Many companies use incomplete data, which leads to incorrect answers, and people do not trust the results. The right way to do it is: first, you get the data; then, you build the infrastructure for the data; then, you make the models; and finally, you get value for the business from the data, the models, and the infrastructure that Snowflake and Databricks provide for the data.

How is generative AI changing fraud and fraud detection?

Generative AI is changing both offensive and defensive approaches. The problem of fraud is getting worse. Fraud attacks are getting really tricky. People are using things like deepfakes, fake identities and phishing emails that are made by computers. These things are becoming very popular fast. Companies, like Sensity AI are keeping an eye on the risks that come with deepfakes. They want to know what kind of trouble deepfakes can cause.

In the security arena, organizations are moving toward adaptive systems. Instead of static rules, they rely on behavioral analysis and real-time detection. Darktrace is an example of this approach. We are entering a phase where it is essentially a battle between AI and AI. Static, rule-based systems are no longer sufficient.

How do institutions balance AI-driven credit decisions with regulation?

There is a tension between being efficient and regulations. Artificial intelligence helps with predictions, and regulations require transparency. 

Institutions implement a mix of strategies and approaches. Artificial intelligence generates forecasts. Then they add more layers to explain things. For example, Upstart works closely with regulators to ensure the explainability of its models.

Human control will always be on the first plan, especially for high-stakes decisions. The future is for AI-enhanced systems that are efficient and auditable.  

What separates institutions that successfully adopt AI from those that struggle?

Shortly, institutions that have successfully integrated and adapted AI separate from others in their architecturly.

Successful institutions are cloud-native, modular, and data-driven. Nubank and similar companies are good examples,  they were built with modern architecture from the very beginning. You cannot implement intelligence into a system that was not designed to support it.

Looking ahead, will AI automate processes or reshape financial products?

From a short-term perspective, AI will reduce costs and advance risk management, and of course, automate the process.  In the long term, it will transform financial products.

We will see more dynamic and personalized financial services. 

One of the most notable shifts will take place in how decisions are made within organizations. We will move from static reporting to interactive, real-time intelligent systems. In finance, speed of decision-making is a competitive advantage, and AI will fundamentally change this. 

Thank you for the great interview, readers wanting to learn more about this fintech accelerator should visit INSART.