Spyros Retzekas, CEO of Qualco Technology and Deputy CEO of Qualco Group, discusses responsible AI adoption in financial services and Qualco Technology’s strategic evolution into an AI-first company in an interview with NPL Confidential.
Qualco Technology launched both the Data, ML & Agentic AI Suite and the Agenly platform within a matter of months. What is the strategic vision behind these moves, and where do you want the company to be in five years' time?
Investing in AI is a deliberate strategic choice. The market is moving towards smarter, faster and more automated operations, and we intend to lead that shift. But it is also in our DNA: listening to what the market needs and responding to it is how we have operated for 25 years. And that experience matters. Successful AI deployment in highly regulated environments requires both domain expertise and advanced AI capabilities. Very few companies have both. We do.
What we are seeing right now is that organisations want optionality. Alongside large transformation programmes, there is real and growing demand for solutions that are modular, fast to deploy and that demonstrate value quickly. Even institutions running major programmes often want to start with something focused and expand from there.
"Investing in AI is a deliberate strategic choice"
We have built our product strategy to serve that reality, and our Data, ML & Agentic AI Suite is where that strategy takes shape: data, intelligence and execution brought together into one continuous operational flow, with AI embedded into how outcomes are actually delivered. Agenly is one expression of that thinking: a cloud-native SaaS platform that helps financial institutions, servicers and any organisation extending credit, including utilities and energy providers, engage customers digitally, guide them towards resolution and enable payments through controlled, scalable interactions.
Where do we want to be in five years? Serving a significantly larger share of the global market, with AI as the engine behind most of what we do. We have the domain knowledge. We have the product foundation. The next step is execution.
AI has become one of the dominant conversations in the financial sector. How mature is the market today in its adoption of AI for financial resolution, and what are the main barriers still holding back wider uptake?
The honest answer is partly. The conversation about AI in financial services has never been louder. However, there is still a gap between interest and implementation, and that gap is worth understanding.
The strongest momentum is where it has always been in this industry: where speed and accuracy matter most, credit decisions, risk assessment, customer operations. Those are the areas where AI delivers measurable impact quickly, and adoption is moving fastest.
The barriers that remain are real. Legacy systems create data silos that make it genuinely hard to move from insight to automated action. And in regulated environments, there is a legitimate caution about trusting AI with customer-facing interactions. That caution is not irrational, it just needs to be answered with the right architecture and governance, not ignored.
What has changed significantly is the investment threshold. SaaS-based AI no longer requires the infrastructure commitment that held many institutions back. That is accelerating adoption across the market in a way we have not seen before.
The organisations that will pull ahead are those that solve the data-to-action problem; that move beyond analytics and into automated, governed execution. That is exactly where the market is heading, and where we are focused.
Many companies are positioning AI solutions for customer engagement. What is the substantive difference between a traditional chatbot and an agentic AI platform such as Agenly, and what problems does it actually solve in practice?
The difference is structural, not “cosmetic”. A chatbot responds to inputs. Agenly is designed to move cases towards financial resolution, and that requires something fundamentally different under the hood.
It starts before the conversation even begins: ML routes each case to the most suitable campaign, channel and moment in time. When the AI agent then engages the customer, through SMS, Viber or WhatsApp, it understands intent, guides the interaction, and presents options. At every point, repayment options, eligibility and treatment strategies are governed by the institution's own business rules. And if a customer wants to speak to a human, the platform honours that immediately.
"A chatbot responds to inputs.
Agenly is designed to move cases towards financial resolution"
For our clients, what this means in practice is straightforward: the high-volume, early-stage caseload, the interactions that consume resources and yield diminishing returns through traditional channels, can be handled digitally, at scale, around the clock. Customers engage on their own terms, through channels they already use, at a time that suits them.
The outcome is faster engagement, lower operating costs and better recovery results. But the point I always come back to is this: Agenly is not replacing judgement with automation. It automates the right parts of the process while keeping financial decisions governed, auditable, and firmly under institutional control.
In a heavily regulated sector such as loan and receivables management, how do you ensure that AI operates with full transparency and in compliance with the European framework, particularly following the implementation of the EU AI Act?
The starting point for us is simple: AI in receivables management cannot operate as a black box. In a regulated environment, institutions need to know what the system did, why it did it, and where human oversight applies. If you cannot answer those questions clearly, you should not be deploying AI in customer interactions.
That is why governance is built into Agenly's architecture from the ground up. The AI manages the interaction: understanding intent, guiding the conversation, presenting options. But it does not make autonomous financial decisions. Repayment options, eligibility criteria and treatment strategies are governed by the institution's business rules. That distinction matters enormously under the EU AI Act, and it is what allows Agenly to operate as a Limited Risk AI system within the European framework. But for us, transparency is not just about meeting a regulatory bar. It is fundamental to building trust with customers. If people do not understand they are engaging with AI, you have already lost.
Traceability is implemented at a granular level. Every conversation, case action and AI-assisted interaction is logged and available for reconstruction. For cases involving vulnerability or complexity, the system escalates to a human agent. We are not trying to automate everything, but automate the right things, responsibly.
"AI in receivables management cannot operate as a black box"
Qualco Technology has spent decades operating in regulated financial environments. Compliance is not something we are learning now, it is embedded in how we build. And that is not a small thing when you are asking institutions to trust AI with their customers.
Qualco Technology operates in more than 30 countries. How does Greece compare with other European markets in terms of the digital transformation of receivables management? Are we ahead of the curve or still catching up?
I would push back gently on the framing of "ahead or behind"; it is more nuanced than that. Greece is at a different stage of the journey, shaped by a very specific decade of NPL management that few other European markets have experienced at the same scale.
Markets like the UK and parts of Western Europe are further along with deploying digital-first engagement at scale. That is partly because competitive pressure arrived earlier, partly because consumer digital behaviour matured faster, and partly because outsourced collections operations have a longer track record there.
The Middle East, interestingly, is moving faster than many expect. There is significant investment in financial infrastructure and a genuine appetite for AI-led solutions that in some respects outpaces more established European markets.
Greece's opportunity is that the conditions are converging right now. Portfolio management has matured post-crisis, the regulatory framework is clearer, and institutional appetite for operational efficiency is higher than ever. That combination creates a real window and having seen what works across a number of countries, we are well placed to help Greek institutions use it to leapfrog a generation of incremental change rather than retrace every step other markets have already taken.
How do you expect the relationship between creditors and customers to evolve over the next five years? Will we see phone-based contact replaced by digital AI agents, and what role will human judgement play in that new environment?
The relationship between creditors and customers is going to shift in a way that I think most people in this industry already sense but have not yet fully reckoned with. It will move from reactive and adversarial to proactive, digital and increasingly self-directed. Customers will expect to manage their situation on their own terms, through channels they already use, at a time that suits them, rather than waiting for a call they did not want to receive.
Phone-based contact will not extinct. But it will be reserved for what it is genuinely suited to: complex, sensitive cases where human presence adds real value. The high-volume, early-stage workload will be handled digitally. That is not a prediction; it is already happening in the markets that moved first.
The human role changes as a result, and I think that change is ultimately a positive one: from transactional work, making calls, logging outcomes, to supervisory and exception-based work, exercising judgement on the cases that genuinely require it. That is a better use of skilled people.
"The high-volume, early-stage workload will be handled digitally"
The organisations that get there first and do it compliantly will build a cost and performance advantage that compounds over time. That window is open now, but it will not stay open indefinitely.
Have we moved from the era of call centres to the era of AI agents?
Yes, but the answer deserves more than a yes.
We have moved into the era of AI agents. But not all AI agents are the same, and that distinction will matter enormously. There is a real difference between a chatbot relabelled as an AI agent and a system that is genuinely agentic; one that understands context, predicts the right moment to act, guides an interaction towards a defined outcome, and does all of that within a governed, auditable framework. One answers questions. The other drives results.
"We have moved into the era of AI agents"
The call centre is not going anywhere. But its role has fundamentally changed. It is no longer the primary channel for high-volume customer engagement, but the escalation layer, the place where human presence genuinely adds something that automation cannot replicate.
The companies that will define the next era in financial services are not necessarily those with the most sophisticated AI. They are the ones that can deploy it responsibly, at scale, within the constraints of regulated environments, which requires domain expertise alongside technology. You cannot shortcut 25 years of understanding how these institutions work, what regulators expect, and where the real operational complexity lies. That is the foundation on which everything we do is built.
This inverview was originally published by NPL Confidential.
See how AI Agents turn conversations into payments