A recent survey of 40+ developers reveals that AI coding assistants are exacerbating communication friction rather than solving development bottlenecks. The Atlassian 2025 report shows that while AI saves development time, it increases time spent on review, rework, and realignment tasks by a comparable amount. Communication latency and fragmented documentation emerged as primary pain points.

Cross-functional alignment proves challenging when 70% of technical constraints need to reach stakeholders who don't regularly interact with codebases. Survey data shows one-third of constraints surface during planning sessions, but articulating complex technical dependencies and translating them into business impact remains difficult. Documentation practices are fragmented, with 52% sharing constraints via Slack and 25% relying on verbal communication with no written record.

AI coding agents compound these issues by being designed to be accommodating rather than challenging. They lack the context and authority to push back on requirements, unlike human developers who engage in productive conflict. Cursor's autonomous coding experiments revealed that agents need human intervention to understand holistic requirements. The core problem isn't AI's inability to write good code, but its inability to refuse writing bad code without proper context. Developers need ammunition for cross-functional alignment, not just better code analysis tools.