A recent tech meetup at Essent's offices in Den Bosch brought together experts to discuss the current state and future of agentic AI systems in software engineering. The discussion highlighted the "Wild West" nature of the field, characterized by rapid changes, uncertainty, and a lack of established best practices. Quality assurance for AI systems poses unique challenges, as traditional testing methods struggle with non-deterministic outputs, necessitating new strategies like evaluations and guardrails. Participants agreed on the importance of selectively delegating tasks to AI systems while maintaining control over critical functions through deterministic guardrails. The conversation also emphasized the need for strong foundational software engineering practices to support AI implementation, as many enterprises face infrastructure readiness gaps that hinder effective AI deployment. The transition from "prompt engineering" to "context engineering" was noted as a shift in focus to improve AI outcomes by providing precise context. The economic implications of running LLMs and the strategic importance of laying solid foundations for AI adoption were also discussed, with an urgency for organizations to act promptly to secure competitive advantages.