Are AI Agents Replacing Jobs? Experts Debate the Real Business Impact | priint:day 2026

What separates Generative AI from Agentic AI, and what does that shift mean for companies, employees, data teams, and business leaders? Moderated by Sebastian Hardung of priint, this panel brings together Katrin Gillett from Adobe, Michael Streit, and Michael Fieg from Accenture to discuss the move from AI experimentation to autonomous, connected, and increasingly productive agentic workflows. The conversation begins with a practical distinction. Generative AI responds to a prompt and creates an output. Agentic AI is designed to understand intent, plan actions, orchestrate tools or other agents, and work toward a defined outcome. The panel explores how learning, memory, autonomy, and process orchestration change the way companies think about automation, Artificial Intelligence, and the future workforce. A central theme is that the biggest obstacle is often not the technology itself, but the quality of the underlying data and the choice of the right use case. Data teams may raise concerns because product data, business rules, and process information are incomplete or inconsistent. The panel argues that these concerns are valuable: a poor process and unreliable data will not become a successful agentic process simply by adding AI. Clear priorities, measurable impact, reliable data, and a realistic business case are essential starting points. The speakers also address the human side of transformation. Employees may fear that their roles will become unnecessary, while organizations may fear falling behind the market. Rather than dismissing these concerns, the panel emphasizes Change Management, honest communication, leadership responsibility, and the importance of involving experienced employees. People who understand existing processes often reveal the limitations and risks that need to be solved before innovation can create real value. Governance is another key issue. Not every process should be fully autonomous, especially where auditability, accountability, regulatory requirements, or clearly documented decision paths matter. The panel discusses the importance of distinguishing deterministic tasks from areas where agentic systems can create additional value. In large-scale procurement and service management, even small efficiency gains can have a significant business impact when processes involve many employees and transactions. The discussion also turns to the economics of AI agents. Token consumption, context windows, model pricing, and infrastructure costs can change the profitability of an AI workflow. Businesses therefore need to monitor the cost per process run, optimize prompts and context, and understand how costs scale. At the same time, the speakers caution against allowing short-term cost concerns to prevent meaningful experimentation with a technology that may reshape entire operating models. To reduce dependency on individual AI models or vendors, the panel recommends open and flexible infrastructures. Interoperable protocols, replaceable models, resilient data platforms, and sensible risk management can help companies adapt as the market consolidates. The panel also considers the future of personal AI agents, from scheduling appointments to handling routine tasks, while recognizing that technical feasibility does not automatically mean people will want to delegate every part of their lives. 🌐 https://www.priint.com/ Read everything about the pr(i)int and publishing universe Blog 🇩🇪 https://www.priint.com/de/blog Blog 🇺🇸 https://www.priint.com/en/blog Blog 🇫🇷 https://www.priint.com/fr/blog Follow us on Social Media LinkedIn 👉   / priint-group   Instagram 👉   / priintgroup   Facebook 👉   / priint.group