ABBYY’s recent AI Summit in Vienna brought together global experts, partners, and customers for a two-day deep dive into the future of intelligent automation. From revealing how fraudsters are weaponizing AI to sharing how ABBYY’s own technology is evolving with LLMs, the event painted a vivid picture of where the industry is headed.
The first day focused on the importance of embracing ABBYY’s partner ecosystem — exploring how shared innovation and integrated solutions are driving transformation for customers around the world. The second day focused on customers, highlighting how ABBYY’s technology is delivering measurable impact across industries. Here are a few of the highlights.
Jump to:
The rise of AI-generated document fraud
The state of AI in 2025: A glimpse into the future
Building a winning AI ecosystem
The rise of AI-generated document fraud
One of the discussions centered around the growing threat of AI-generated document fraud with Jan Syrinek, Head of Product from Resistant AI. Syrinek shed light on the growing scale and sophistication of document fraud, highlighting how AI tools and “template farms” are enabling fraudsters to generate highly realistic fakes with alarming ease.
Resistant AI has responded by investing heavily in threat intelligence, analyzing over 200,000 fraudulent templates sourced from more than 180 countries. Their detection methods go beyond surface-level analysis, using neural networks to spot regional inconsistencies, watermarks, and subtle behavioral patterns.
According to Syrinek, fraud prevention was once a low priority for some businesses, but that’s changing fast. More companies now recognize the value of integrating advanced fraud detection into their operations—not just to catch obvious fakes, but to stay ahead of rapidly evolving threats.
Examples shared included fake Chase bank statements, Chinese licenses, and tax-accurate New York receipts—all created using AI. The message was clear: the threat is real, global, and growing.
Learn how LLMs are versatile, enabling business to use them effectively in an increasingly competitive marketplace.
The state of AI in 2025: A glimpse into the future
In a panel featuring industry experts, the state of AI in 2025 was described as "chaotic" by Christina Duta, Senior Director of DevOps and Intelligent Automation at Fortrea, citing wide disparities in AI maturity across industries. While highly regulated sectors remain cautious, AI is becoming central to technology strategy. The panel emphasized the need to understand the "why" before adopting new technologies, with many executives disconnected from the on-the-ground realities, particularly in industries like construction.
Bernhard Hecker, Enterprise Transformation & Automation Strategist, sees 2024 as the year of retrieval and generation, with 2025 marking the rise of agentic AI. Meanwhile, Alexander Hanff, Cybersecurity and AI Governance Expert, Hanff & Co AB, pointed to "cautious opportunity" amidst a fragmented regulatory landscape. As AI systems grow more autonomous, experts stress that AI’s true potential lies in data quality, human judgment, and proactive regulatory compliance.
Explore enterprise-scale document automation, powered by AI, such as ABBYY’s FlexiCapture.
Building a winning AI ecosystem
“How many of you believe you have a successful AI ecosystem in your organization?”
That was the opening question from Cristina Duta, Senior Director of DevOps and Intelligent Automation at Fortrea, during a panel with Nigel Houghton, Director of Strategic Partnerships at M-Files, Ina Krebs Partner & General Manager DACH, Nordics. Managing Director UK at Novelis, Tomasz Gradzik, Director of Sales / Chief Executive Officer (CEO) at FORESIGHT, Alicja Wolanczyk, Partner Strategy and Alliances at ABBYY, and Jeremy Burgess, Director at Workflow Doctor Ltd.
The discussion focused on what it takes to build an AI ecosystem that works. Speakers described it as a value chain: people → data → technology → outcomes. At the center is strategy.
Jeremy Burgess noted that tools like ABBYY’s FlexiCapture can support many use cases, but only when applied with clear intent. He shared a hospital project where the original goal was to reach 60% straight-through processing (STP). That target led to new issues. The root cause was inaccurate procurement data.
The team changed the approach. Instead of measuring STP, they focused on effort saved. AI was used to match invoice line items—applying logic similar to human judgment rather than extracting data.
The main point: success depends on strategy, alignment, and collaboration. AI should solve defined problems, not serve as the starting point. Process, people, and purpose must guide every step.
Learn how intelligent automation enables business transformation.
Navigating the legal and ethical landscape for AI in HR
Alexander Hanff discussed about how deploying AI in HR comes with complex legal and ethical challenges. GDPR Article 22 restricts automated decision-making without explicit consent, especially in recruitment, and consent isn’t feasible in employment due to the employer-employee power imbalance. Ethical AI use is essential, with long-term risk mitigation taking precedence over short-term ROI.
Lack of transparency in AI systems undermines accountability, as seen with unsupervised AI causing harm. To ensure compliance, engage legal teams early, using tools like Data Protection Impact Assessments (DPIA). Address AI model biases, even in seemingly neutral designs. Responsible AI not only ensures legal compliance but can also provide a competitive edge, making governance a strategic asset for businesses.
Explore the ABBYY Trust Center to learn about our security posture and request access to our security documentation.
AI implementation in corporate environments
Bernhard Hecker discussed the key considerations for AI adoption in corporate settings. While innovation is vital, Hecker stresses that AI must create tangible value for organizations—experimenting for the sake of it doesn’t cut it in a business environment.
Real AI integration requires more than just technology—it involves aligning company culture, understanding internal processes, and ensuring organizational readiness. Hecker warns against jumping into AI without a clear strategy, highlighting that executive pressure to "do something with AI" often leads to failed projects.
Hecker highlighted failures like McDonald's ChatGPT-driven drive-thru, where AI misunderstood orders and a US law firm using ChatGPT, which invented fake court cases, leading to case dismissals. These failures underline the importance of context, control, and validation in AI use.
He also shared a success story: when the U.S. government released 80,000 JFK-related files (mostly unstructured PDFs), ABBYY’s document intelligence API was used to convert them into structured data. This allowed for advanced querying, such as answering "Who killed JFK?"—confirmed as Lee Harvey Oswald.
Additionally, 80% of enterprise data is unstructured, like documents and emails. Organizing this data is crucial for AI success. AI must also be transparent and explainable to avoid becoming a “black box” in critical decision-making.
As the ABBYY AI Summit Vienna wrapped up, one thing became clear: the future of intelligent document processing will be shaped by a balance of innovation, responsibility, and real-world impact. From decoding the anatomy of document fraud to exploring the ethical and strategic guardrails for enterprise AI, the event showcased how purpose-built AI delivering over 90% straight-through processing, 99%+ accuracy, and thousands of hours saved each month, ABBYY is proving that AI isn’t just a vision—it’s a results-driven reality. As organizations face increasing pressure to digitize smarter, faster, and more securely, gatherings like this Summit offer both the clarity and collaboration needed to move forward with confidence.
For more insights visit the full ABBYY AI Summit series page. Register now to attend the ABBYY AI Summit London on the 21 May.