Key facts
- Professionals are increasingly spending more time managing AI agents and reviewing AI-generated work.
- This managerial role involves setting expectations, delegating tasks, evaluating performance, and intervening when AI systems err.
- Companies like Salesforce are developing AI agents designed to perform tasks autonomously, requiring human oversight.
- Wipro has achieved significant productivity gains through AI, with engineers managing multiple AI agents.
- A survey by Boston Consulting Group indicates 47% of employees spend more time managing AI than doing their core work.
- AI is taking over simpler tasks, leaving humans with more complex work, judgment, responsibility, and oversight.
Professionals across various fields are increasingly finding themselves in managerial roles overseeing AI systems, rather than focusing on their core craft. This shift, driven by the widespread adoption of AI, requires individuals to set expectations for AI agents, delegate tasks, evaluate their performance, and take responsibility for the final output.
Companies are actively building products and strategies around this evolving work dynamic. Salesforce, for instance, has introduced "job-ready agents" for sales, customer service, and operations, enabling human employees to train and coordinate these AI tools. Similarly, Wipro, an Indian IT services company, has leveraged AI to achieve productivity gains equivalent to 20,000 workers, with its chief technology officer envisioning engineers managing multiple AI agents.
For some, this new role is seen as a form of "workplace flex," where they "rule over AI" rather than being replaced by it. However, many professionals, such as software engineers who opted for individual contributor tracks, find themselves de facto managers of machines without the expected career advancement or compensation. Sinda Khenine, a software and AI engineer, noted that managing AI systems requires significant effort in coordination, model selection, context provision, and output review, often overshadowing the original engineering problem.
This trend extends beyond the tech industry. Patricia Diaz-Hymes, a product marketing consultant, estimates spending about 30% of her time supervising AI, likening it to management by giving agents names and specific jobs. A survey by Boston Consulting Group found that 47% of employees globally spend more time managing AI than performing their own tasks, with AI handling simpler duties and leaving humans with more complex oversight and responsibility.