AI's Impact on Tech Roles Shifts to Supervision
AI is altering the nature of technical work, shifting the focus from writing code to supervising and orchestrating AI agents, as seen at companies like Canva. This trend aligns with broader findings that AI's employment impact is more evolutionary than revolutionary, with IBM now tripling its hiring for AI-related roles after previously announcing plans to replace jobs with AI.
- IBM is re-engineering its entry-level positions to focus less on tasks that can be automated and more on human-centric responsibilities like client engagement and collaboration. As part of this shift, the company has rewritten job descriptions to reduce the emphasis on technical work that AI can perform. - This strategic change aligns with broader executive sentiment; in a 2025 Forbes survey of over 1,000 C-suite leaders, 59% stated they believe AI will ultimately create more jobs than it eliminates. Furthermore, 94% of these executives predict that fewer than 5% of jobs will be eliminated over the next two years due to AI. - The role of a software developer is evolving from a pure coder to an "agent operator" or "system designer" who guides AI tools. This involves a greater focus on high-level system architecture, prompt engineering, and ensuring the security and scalability of AI-generated software. - At Canva, the integration of AI agents is intended to have them function as "proper teammates" for design, marketing, and sales tasks. Co-founder Cameron Adams indicated these agents would assist with research, data pulling for marketing strategies, and tidying up presentations. - The demand for AI proficiency is now a factor in hiring, with Canva requiring applicants for many software development roles to demonstrate skill with AI-assisted coding tools during technical interviews. This reflects a growing industry trend, as Dropbox's Chief People Officer, Melanie Rosenwasser, noted that younger employees are significantly more proficient with AI tools. - In the HR and compensation field, AI is being used to analyze large datasets for salary benchmarking, identify pay disparities, and ensure equitable pay structures. AI-powered tools can reduce payroll errors by up to 90% and cut down on the time spent on compensation analysis by 40%. - Agentic AI workflows are being built using specific, repeatable design patterns such as "Reason and Act" (ReAct), where an agent loops through reasoning, acting, and observing. Other patterns include multi-agent collaboration and incorporating a human-in-the-loop for oversight. - To ensure user trust and control in agentic systems, UX design patterns are emerging, such as "Intent Preview," which defines the AI's plan before action, and "Action Audit & Undo," which provides a safety net for potential errors.
