At this AI startup, the engineers became the bosses, and agents do the coding

In a groundbreaking shift within the AI landscape, Wayfound.ai has redefined the roles of its engineering team, promoting them to managerial positions while AI agents handle coding tasks. Tatyana Mamut, CEO and co-founder of the startup, reveals that her streamlined team of just two engineers significantly outperforms larger teams, showcasing higher productivity and faster feature rollout than her previous experience at Amazon Web Services in 2017. Mamut explains that the workforce now primarily consists of engineering managers who oversee AI agents capable of coding, allowing for a more efficient process. The engineers not only manage these agents but also engage directly with customers to tailor features to their needs. This innovative approach began in 2024, initially utilizing tools like ChatGPT, but has since evolved to favor Claude Code for its enhanced capabilities. The operational efficiency is striking; coding agents have eliminated the need for extensive testing periods previously required, drastically reducing the project timeline. Engineers collaborate in focused meetings, prioritizing tasks and requirements in a matter of minutes, streamlining the workflow significantly. Financially, the startup has benefitted as well. By enabling engineers to take part in customer interactions, Wayfound has avoided the need for additional customer success personnel. Mamut emphasizes that this transformation allows engineers to engage more with clients, enhancing their role from mere coders to versatile 'builders'—a trend seen in major tech firms. However, Mamut cautions against the potential pitfalls of relying solely on AI agents without proper oversight. She warns that neglecting continuous training and supervision could lead to serious operational issues. The rise of AI tools presents an opportunity for growth, but companies must remain vigilant in their application and integration. Mamut also critiques the traditional Software-as-a-Service (SaaS) model, arguing that it must evolve to incorporate AI agents effectively. As companies like Salesforce face scrutiny over their stock performance amidst AI advancements, she suggests that adapting to become more 'agentic' is crucial for survival. Cost-cutting measures, particularly in workforce size, are expected as firms pivot towards this new model. In the face of these changes, Mamut asserts that human roles will remain essential for managing AI agents and maintaining authentic customer relationships. “People want to buy from other people,” she states, highlighting the ongoing need for trust and connection in business interactions.