More than a century ago, Ford’s moving assembly line reorchestrated work. Instead of a master mechanic walking to each car to perform complex tasks, the car moved to workers, who each executed a single repeatable action. As a result, work became easier, costs fell and return on investment skyrocketed.
Here are two areas in which this might unfold:
• Onboarding. In a large enterprise, bringing a new hire up to speed requires 10 to 20 hours of effort. Forms, policy explanations and training sessions add up quickly. In an AI-first model, an onboarding orchestrator bot coordinates with AI agents that handle compliance (background checks, payroll setup, training) and equipment (granting access to systems and tools). A virtual assistant answers questions from the new hire.
• Software delivery. Today, new information-technology features—such as launching a new process control program—are specified by operations managers, designed by user-experience teams, coded by developers and tested before use. This cycle can take 80 to 85 business days, with error rates around 10% to 15%.
In an AI-first setting, a digital IT orchestrator bot coordinates AI agents that do everything from product specifications and design to coding and testing, delivering the first version. People then review the newly created features and test their AI-generated code and design. Instead of arriving in 3 months, features are ready in weeks. Time to delivery falls by 60% to 70%, while code quality improves through continuous AI-driven testing.
The common threads of AI-native work are clear: Orchestration moves from humans to AI. Specialized AI bots handle repeatable tasks. Human experts intervene when judgment, negotiation or oversight is required. Results come faster, with lower unit costs and better customer experience.
Henry Ford pioneered a new way of doing business in the 1900s. We have a similar opportunity in 2025. While Ford’s assembly line turned employees into specialists, putting AI at the center of business turns us all into generalists, allowing us to be creative, enlist problem-solving skills and handle ambiguous work.
Classroom discussion questions:
- How else might AI be used in factories?
- What is an AI “agent”?
Greetings from the National Autonomous University of Mexico. We discussed this news in the Process Strategies and Product and Service Design group, and I’m sharing a summary of our discussion:
The implementation of artificial intelligence in companies involves a complex process that requires investment in training and work hours, as well as a transformation of existing processes within the organization. It is necessary to begin preparing employees for the use of these technologies, even if this means redefining established structures and dynamics. At the same time, the adoption of AI generates the need for new professional profiles specialized in its use, which may lead to an increase in the unemployment rate, as many tasks can be resolved more quickly and efficiently than by specialists with advanced training. The risk of placing artificial intelligence at the center of strategic decision-making must also be considered, as excessive technological dependence can make decisions of great importance to the company. Finally, in developing countries, implementing AI may not be a viable option in the short term due to economic and infrastructure limitations, which delays the possibility of integrating it into their production processes.
César Oliva, EPDPS, UNAM.