AI - your new advantage
Or a new risk?
25.09.2025
AI - your new advantage
Or a new risk?
25.09.2025

#ISCG_AI2025 conference behind us, it was a day full of inspiration
and practical knowledge in the heart of Microsoft Poland.
We then moved on to key AI security issues, pointing out the complex threats that can surprise companies unprepared for the new era of cyberattacks, to AI and ML demonstrations with Azure Machine Learning and Cosmos DB that make real-time fraud detection not only possible but also enhance decision-making security.
A pill of valuable knowledge, complemented by the principles of responsible implementation of AI in business, legal meanderings of GenAI, compliance with the AI Act and practical case studies in the work of a lawyer.
Finally, we took a journey through the evolution of artificial intelligence toward Agentic AI - intelligent systems that can function without direct human intervention.
Thank you for the exchange of experiences, the knowledge imparted and the incredible energy of the event to be able to look at digital transformation from a new perspective.
Thematic sessions
Connectory and Governance in Practice
In this session, you'll learn about the practical applications of agents in Microsoft Copilot, from creating them, to integrating them with your organization's systems using Copilot Connectors, to effectively managing their performance within governance. You will learn how to prepare your organization for the deployment of agents, how to take care of security, compliance and control of their activity, and how to use their potential in different departments of your company. The session includes demos, specific deployment scenarios and tips you can apply right away.
Microsoft 365 Copilot is a powerful tool that, when used appropriately, can significantly increase productivity in an organization... but only if it is implemented consciously and securely. Many companies make basic mistakes: ignoring data quality, overlooking security aspects, failing to plan implementation or prepare users for the new way of working. The effects are familiar - instead of support, often chaos and risk.
This session will show how to prepare for a Microsoft 365 Copilot implementation with a head start: from reviewing and organizing data, to securing information, to planning the change and educating users. The most common pitfalls will also be discussed, and how to avoid them.
Session Plan:
- key elements before implementing Copilot
- security, permissions and data quality
- lessons learned from deployments and the most common mistakes
AI on target - what every board should know
An expert with many years of experience in the Defense Ministry and the design and security of information systems, he will present why artificial intelligence has become a new target for attacks - including in the private sector. He will present a typology of the most dangerous threats, such as adversarial manipulation, training data poisoning, model theft and prompt injection, which can lead to loss of data, reputation and competitive advantage. He will show that AI without proper safeguards can be easily manipulated - even without the knowledge of the end user. He will emphasize the need to build awareness and implement protection mechanisms at the enterprise level, and point out regulatory gaps that can cost companies more than they expect.
With AI and ML - see how technology protects your business from fraud
During the session, we will demonstrate how modern technologies - Azure Machine Learning, Event Hub and Cosmos DB - enable real-time fraud detection. We will show a complete, ready-to-implement scenario in which artificial intelligence identifies suspicious transactions before they cause damage. Participants will see how AI and automation support quick operational decisions, increasing security and reducing losses - without the need to build complex infrastructure.
Dr.
Aleksandra Auleytner
Partner
Head of DZP's IP&TMT Practice
Chief Information Officer DZP
Counsel Aleksandra Auleytner and Counsel Marek Laskowski, Chief Information Officer
GenAI and the law - how to safely implement artificial intelligence in business operations
Experts from the law firm Domański Zakrzewski Palinka (DZP) will focus on the key legal aspects of using generative artificial intelligence in business - with a particular focus on the challenges posed by EU regulations, including the AI Act. Participants will learn practical tips for the compliant and secure implementation of GenAI solutions in organizations. Speakers will also present case studies in which GenAI supports document analysis, legal process automation and legal risk management. Special emphasis will be placed on data protection issues, liability for AI-generated content, and technology due diligence. An example of how GenAI is used in the practice of DZP's law firm - in working on transactional documentation and legal analysis - will also be presented.
In a world full of agents. Are we heading toward a future designed by AI?
During the meeting, we will examine the evolution of artificial intelligence models - from large language models to advanced AI systems capable of independent decision-making and reasoning. We will see how combining these models with reinforcement learning leads to fully autonomous agents that operate without direct human intervention. Special attention will be paid to so-called Agentic AI, or intelligent systems that can analyze situations, make decisions, and even initiate and execute complex tasks in a dynamically changing business environment. The aim of the presentation is not only to show current trends, but also to inspire participants to make practical use of the potential of agentic AI as one of the key factors in enterprise transformation
In Copilot Studio - The New Generation of Automation
Copilot Studio opens a new chapter in task and business process automation with autonomous AI agents. These intelligent automatons can independently plan and execute actions, make decisions based on data and even communicate with other agents - without the need for constant human supervision. But does it actually work? Let's check!
During the session, I will show how to build and deploy an autonomous agent in Copilot Studio. Based on a real-world scenario, I will demonstrate how the agent uses knowledge sources to communicate with the user, how complex tasks can be delegated to it, and its ability to work with other agents. I will also show how effectively such an agent performs in a production environment.
Session Plan:
- How autonomous agents work and what distinguishes them from regular agents
- designing an autonomous agent at Copilot Studio
- What works and what doesn't yet - using a real-life agent as an example
