This webinar reveals why AI "Agents" act like brilliant but unpredictable interns who need strict supervision to work in the real world. Using an SAP case study, we will show you how to move past the hype by building agents with "manuals" (pseudocode) and safety nets to catch their creative mistakes. You will learn why you can’t rush these projects and must build them piece-by-piece to keep the AI from getting confused. Ultimately, you'll gain a practical blueprint for turning erratic AI behaviour into a reliable, professional tool.
Agenda for the session
- Understand why AI agents require supervision
- SAP Case Study: Building reliable agents step by step
- Designing effective controls: manuals and safety nets
- Live Q&A with the speaker
About Speakers
Vishnu Subramanian
Associate Manager - AI/ML Computational Science Accenture
An Associate Manager in AI Engineering at Accenture with 5+ years of experience in Data Science and AI. He specializes in building Generative AI and LLM-based solutions, including Agentic workflows and GraphRAG systems. Previously a Lead Data Scientist at Great Learning, he contributed to AI-driven product development and GenAI programs, enabling efficiency and scalable adoption. He holds an MS in Data Science from Heriot-Watt University and an MS in International Business from Hult International Business School.
