Slavena Hristova, Director of Product Marketing, ABBYY

Today we're delighted to speak with Slavena Hristova, Director of Product Marketing, Document AI at ABBYY. With over 22 years at ABBYY, Slavena brings deep expertise in how AI-powered automation is transforming the way financial institutions work with documents and data - particularly as generative AI makes document fraud faster, cheaper, and more accessible than ever before.

My questions are in bold - over to you Slavena:

Who are you and what's your background?

I've spent my entire career - just over 22 years - at ABBYY. I'm a lifer!

Born and raised in Russe, Bulgaria, a beautiful city characterised by Baroque and Rococo architecture on the Danube river, I came to Germany for my studies in Business Administration, which also led me to ABBYY.

I joined the company as part of my mandatory study internships and quickly developed a passion for translating complex technologies into language that resonates with non-technical audiences. This led me directly into product marketing, where I've developed my professional experience and accumulated a wealth of industry knowledge and expertise.

What is your role, and can you give us an overview of the business?

I'm Director of Product Marketing, Document AI at ABBYY, and I am passionate about how AI-powered automation can change the way businesses work with documents and data. I help customers work smarter by using technologies like Document AI, OCR, and business process automation to streamline processes, make better decisions, and drive digital transformation.

I'm excited about the future of intelligent automation and how it can help teams spend less time on manual tasks and more time focusing on what really matters.

ABBYY provides Document AI and Process AI solutions to large enterprises all around the globe. It has been interesting to observe and influence how automation and digital transformation has affected and truly revolutionised the customer experience in financial services. And equally fascinating to see how the exact same technological advancements have accelerated financial crime over the past few years.

ABBYY helps financial institutions automate their document-heavy business processes and significantly improve customer experience by feeding accurate data downstream. At the same time we have established partnerships with document forensics providers, who help us make this data not only accurate, but also reliable by adding the checks and balances to catch altered or fraudulent documents before they can impact important decisions.

By bringing together Document AI with AI-powered forensic document checks right at the point where documents enter a system, suspicious activity can be identified quickly. Incoming documents such as IDs, payslips, and invoices, are analysed for signs of manipulation, whether in layout, metadata or completely fake, and a clear fraud risk score is provided before data even gets extracted from the document.

Your work centres on detecting document fraud in financial services. How is AI changing the nature of document fraud right now?

AI is fundamentally changing the nature of document fraud by making document manipulation faster, cheaper, and more accessible. Financial institutions have spent years strengthening fraud detection systems using transaction monitoring, identity verification, and machine learning to identify suspicious behaviour across payments, accounts, and digital channels. While we at ABBYY focus on extracting the data from the documents that drive KYC, onboarding, lending, and many other processes even the most accurate extraction from a fake document still leads to harmful decisions for the organisation downstream. This is why we are advocating for combining fraud detection with document processing so that no harmful data enters the process to begin with.

As organisations become more digital, documents are increasingly submitted through mobile apps, online portals, and email. Generative AI and advanced image editing tools now allow fraudsters to create or alter these documents with a level of speed and realism that was previously difficult to achieve.

The shift is that document fraud is moving from being a largely manual, low-scale activity to a more sophisticated and scalable threat. Fraudsters no longer need advanced technical skills or significant resources to produce convincing fake documentation. Fabricated records can be generated, modified, and deployed within minutes.

This creates a new challenge for organisations: the quality of their fraud detection and risk decisioning depends on the accuracy of the data they receive. If fraudulent documents enter the process at the point of onboarding or assessment, even the most advanced analytics and compliance checks may be operating on the wrong information.

The result is that verifying document authenticity has become a critical foundation for effective financial service operations. As organisations continue to automate decision-making, they need new ways to verify not just the customer, but the evidence those decisions are based on.

In the age of synthetic identities and generative AI, can documents still be trusted? What does verification need to look like now?

In the AI era there are very few things left that can be automatically treated as genuine and trustworthy. Documents make no exception. The tools and scale fraudsters are now using to create convincing fake identities, alter legitimate documents, and generate realistic-looking files can no longer be matched by teams manually checking incoming documents and simply eyeballing whether a document looks correct.

Traditional fraud prevention systems have focused largely on transaction monitoring and behavioural analytics, looking for unusual account activity, payment patterns, or behaviour. While these are still important, they often rely on the assumption that the information entering the system is accurate. If the original documents or identity data have already been manipulated, even the most advanced monitoring tools could be compromised.

A stronger approach starts earlier, right as documents enter the organisation. While intelligent document processing solutions focus on document ingestion, ensuring that documents are properly captured, processed, and converted into usable data, including correcting image issues and extracting information from complex files, document forensics handles document authenticity as an important layer of fraud detection. While specialist fraud detection platforms analyse risk signals and suspicious behaviour, organisations should also verify that the documents entering those systems are genuine. Advanced authenticity analysis can identify signs of manipulation, such as unusual metadata, inconsistent fonts, altered layouts, or hidden digital changes.

Beyond the document itself, a process intelligence layer helps organisations understand how documents flow through their processes, which helps detect anomalies. Fraud patterns often emerge through repeated document submissions, unusual workflow behaviour, or attempts to bypass established processes.

Finally, risk-based decisioning brings together document authenticity checks with other signals, such as identity verification, behavioural analytics, and transaction monitoring. This allows organisations to automate decisions while directing higher-risk cases for human review.

By incorporating document authenticity verification into digital workflows, in combination with reliable document processing organisations can reduce the risk of both inaccurate and manipulated information entering critical systems, creating a stronger foundation for reliable downstream decisions, fraud prevention and compliance processes.

What should banks and financial institutions be doing differently to protect themselves against AI-enabled fraud?

Research consistently shows that institutions with higher levels of automation maturity are more effective at managing operational risk.

Fraud prevention is most effective when document processing, identity verification, and fraud detection solutions work together rather than in isolation. ABBYY's role is to help ensure trusted document data flows into those downstream systems so they can make better-informed decisions.

But technology alone is not enough. The next generation of fraud prevention will require a broader shift in how financial institutions design and operate their core processes.

Many fraud controls today are layered onto legacy workflows that were never designed for automation, real-time decisioning, or AI-driven analysis. Documents, identity checks, and transaction monitoring often sit in disconnected systems, with limited visibility across the full customer journey. As a result, risk signals remain fragmented, and operational teams struggle to respond quickly to emerging fraud patterns.

Addressing modern financial crime therefore requires more than adding new fraud detection tools. It requires organisations to rethink how data moves through their operations.

Institutions must modernise the systems and processes that handle documents, identity verification, and case management so that risk signals can be captured, correlated, and acted upon in real time.

Where do you see the arms race between fraudsters and detection heading over the next few years?

Fraud is unlikely to become less sophisticated in the coming years. Advances in AI and automation are enabling criminals to generate increasingly convincing digital artifacts at scale. Synthetic identities and AI-generated documents are already commonplace in fraud attempts.

At the same time, financial institutions are accelerating their own adoption of AI and automation technologies. Intelligent document processing, automated decision engines, and digital customer journeys are becoming standard components of modern financial services operations.

This convergence creates both opportunity and risk. Automation can dramatically improve operational efficiency and customer experience. But automated systems depend on accurate, trustworthy data.

As a result, the industry will increasingly rely on a combination of technologies - including document authenticity verification, identity verification, behavioural analytics, and specialist fraud detection platforms - to keep pace with evolving threats. Financial institutions that fail to validate the integrity of the documents and data entering their business processes risk automating flawed decisions at scale.

The consequences of this could extend beyond individual businesses, affecting the wider industry and the people who rely on these services.

If you had a magic wand, what one thing would you change in the banking and/or FinTech sector?

Well, I'd say ABBYY's Document AI is pretty close to a magic wand!

If I could change one thing in financial services, it would be to ensure that every decision is built on accurate, trustworthy data.

Banks and insurers process enormous volumes of documents every day. The challenge isn't simply digitising them. It's understanding their content correctly, extracting the right information, and ensuring that the data flowing into downstream systems is complete, reliable, and fit for purpose.

That becomes even more important as organisations embrace AI and autonomous automation. AI agents and intelligent systems can make decisions and execute processes at a scale humans never could. As funny or absurd as some of the AI mistakes we hear about in the news may sometimes seem, they also highlight a serious reality: in financial services, errors are anything but harmless. Mistakes in this domain can have catastrophic consequences - for individuals, for businesses, and for entire institutions. And because AI simply amplifies whatever data it is given, if inaccurate or incomplete information enters the process, it doesn't just create isolated errors; it scales those mistakes faster and further than ever before.

That's why I believe high-quality data is becoming one of the most strategic assets in financial services. With AI-powered Document AI, organisations can transform unstructured documents into trusted, structured data that enables smarter automation, more confident decisions, and AI systems that businesses can actually rely on.

The future of banking won't just be defined by more AI. It will be defined by AI that can make the right decisions because it's built on the right data.

Can you list 3 people you rate in FinTech or financial services that our readers should follow on LinkedIn, and why?

While there are many great voices in FinTech, these are three I consistently come back to because they each offer a different lens on the industry.

Brett King helps take a peek beyond today's headlines. His content focuses on where banking is headed, from AI and embedded finance to digital identity and the future of customer experience. I appreciate that he challenges conventional thinking and encourages readers to consider what financial services could look like five or ten years from now.

Dr. Efi Pylarinou is a go-to source for understanding the broader technology landscape. She has a remarkable ability to connect developments across AI, digital assets, payments, regulation, cloud, and banking technology into a coherent picture. If you've ever struggled to keep up with the pace of innovation in financial services, her content helps separate the meaningful shifts from the noise.

Theodora Lau brings an essential human perspective to every discussion. While many conversations in AI and FinTech focus on technology itself, she consistently reminds us that innovation only matters if it creates trust, inclusion, and better customer outcomes. Her work is a valuable reminder that responsible innovation is just as important as technical innovation.

No single voice has all the answers, but this combination provides a balanced perspective that helps navigate one of the fastest-moving industries today.

Many thanks to Slavena for taking the time to share her insights with FinTech Profile. You can learn more about ABBYY on their website.