Workflow Founder Presents on Real-World AI Limitations at Johns Hopkins Conference
Ali Dasouqi, founder of Workflow, recently presented at the 2025 Business Analytics, AI, and Cherry Blossom Conference hosted by Johns Hopkins University, where he delivered a talk titled “Navigating the Real World: The Limitations of AI in Address Orientation.”
The presentation focused on a critical but often overlooked challenge in artificial intelligence: the gap between impressive model performance in controlled environments and the practical difficulties AI systems still face when operating in messy, real-world spatial and contextual settings. By examining address orientation and spatial reasoning as a case study, the talk highlighted how AI can still struggle with ambiguity, incomplete context, and real-world navigation logic despite rapid progress in broader language and reasoning tasks.
The session reflected Workflow’s broader philosophy that successful AI adoption requires more than enthusiasm around capability — it also demands honest attention to limitations, deployment risks, and the operational realities of using AI in the field. This perspective continues to shape the company’s work across business automation, technical systems, and vertical AI applications.
Presenting at a Johns Hopkins conference also reinforced Workflow’s role at the intersection of applied AI, technical rigor, and real-world implementation. As the AI market matures, conversations like these are becoming increasingly important: not just what AI can do, but where it still breaks, and how those gaps should inform product design and deployment decisions.