AI agents are being explored as a way to help businesses break down internal barriers, known as silos, where different departments or teams operate independently and may not share information or resources effectively. A recent report by Deloitte highlights that the rapid integration of AI and the development of cloud computing are prompting companies to rethink their technological strategies, aiming to improve speed, efficiency, and growth. However, despite these advancements, many professionals and their organizations are still not fully prepared to take full advantage of AI's potential, according to Michael Privat, director of data and engineering at Availity. Privat warns that implementing AI into existing systems can lead to inconsistent results, potentially causing disruptions in applications and processes as they transition between product development and engineering. He compares the current state of IT services to a Pachinko game, where ideas bounce through various layers of an organization, often leading to unpredictable outcomes. He cautions that if AI is not properly integrated, it may only automate poor initial decisions, making the situation worse rather than better. To address these challenges, Privat suggests that organizations should focus on solving real customer problems rather than fixating on the latest technology or products. His company is organized around "market domains," which are teams dedicated to understanding and solving specific issues faced by customers in particular markets. Beneath that, cross-functional teams are designed to work quickly, avoiding unnecessary handoffs or endless meetings, which allows experts to execute tasks efficiently. He compares this approach to giving everyone a high-performance car but limiting their speed to 30 km/h (20 mph), which would be pointless and counterproductive. For professionals looking to improve their AI skills, Privat recommends analyzing decision-making workflows to identify where AI can be effectively applied. He also encourages rethinking how work is defined, treating AI results as a production system and delegating tasks that add little value. He warns that AI may increase workload rather than reduce it and advises viewing AI as a mirror that reflects the quality of existing processes rather than a magical solution to problems. He emphasizes that AI does not fix flawed engineering cultures but instead highlights them clearly. Privat also hosts a weekly forum with key opinion leaders and engineers, discussing topics such as AI factories and the need to rethink AI's role beyond being a powerful tool.