Can AI-enabled automation replace your company’s entire labor force?

How much will technology change affect labor? According to research informing our 2017 research report, “Reshaping Business with Artificial Intelligence,” not as much as one would expect. Less than half of more than 3,000 survey respondents expect AI to reduce their organization’s workforce in the next five years, and just 30% are fearful that AI will automate their own jobs. But as technology relentlessly improves, significant changes may still be on the horizon. To what extent can AI-enabled automation replace the entire labor force of an organization?

Consider these companies:

  • The Moby Mart, a store on wheels, brings groceries to customers in Shanghai. Unlike most other retailers, each mobile store has no staff, not even cashiers.
  • The Aidyia hedge fund consolidates recommendations generated with multiple types of artificial intelligence. Free from the constraints of physical products, algorithms trade digital assets without human involvement.
  • BottleKeeper, a producer of stainless steel containers, has no employees.

These companies, encompassing both physical and purely digital products, showcase the potential of autonomous companies. They are taking the replacement of labor with technology to an extreme. While most organizations are far from this extreme, what can managers in far less autonomous companies learn from — and be wary of — the trend toward autonomous companies?

Moving Remaining Labor Outside the Company

An autonomous hedge fund such as Aidyia’s probably surprises us less than an autonomous retail store. Since information products are already digital, full automation seems more natural.

But Moby Mart and BottleKeeper are decidedly based on physical goods. Modern information and communication technologies enable the automation of most of the companies’ labor. Then, specialized external companies perform any aspects that still require labor (such as manufacturing or delivery). Furthermore, because these external partners work with many other companies, economies of scale enable Moby Mart or BottleKeeper to execute aspects that still require labor better and cheaper than they would get by using internal employees — for example, it is hard to duplicate the reach and efficiency of global delivery companies.

Ask yourself, Do any labor-intensive tasks still need to be internal to your organization?

Trading Fixed and Variable Costs

Even if Moby Mart, Aidyia, and BottleKeeper operate autonomously, they aren’t completely autonomous companies. Instead, these companies made strategic decisions to invest in fixed costs to avoid variable costs. For example, they set up payment systems in mobile stores, developed algorithms and computer systems, and contracted relationships with external partners.

This trade-off isn’t new. Companies have long done this by, for example, investing in a high-quality machine to avoid future downtime or defects. What is changing is the lengths to which companies can take this trade-off. For Aidyia, setting up the AI systems to make recommendations took considerable development effort. Furthermore, detailed specifications required investments in time and resources to prepare for contingencies that may never happen — a loss of welfare and surplus. But the benefit is the minimal cost in ongoing operations.

What variable costs can you avoid through additional up-front investment?

Beware of Hidden Costs

Reducing employees doesn’t necessarily mean that their tasks are no longer performed. They may just shift elsewhere.

For Moby Mart, someone must monitor operations, handle unexpected flat tires, coordinate resupply, assess changing market demands, etc. All these tasks still sound a lot like labor, but if they remain small, the straws may not break the camel’s back. But the organization may be using higher-cost labor to do tasks that could be aggregated and use less labor — reinventing the idea of an employee.

By automating, are you just hiding labor in different places?

Expect the Unexpected

Aidyia trading may perform well now, but how it will adapt when change inevitably becomes necessary isn’t clear yet.

An autonomous organization may not be quite so autonomous as the world changes. Some of the fixed costs investments may not be fully paid off when market changes require reinvestment. Or partner companies may change their business model or interface. Or an autonomous competitor may observe the profits from an automated company, build a better mousetrap, and wring out profits.

How will your autonomous operations cope with change?

The traditional Cobb-Douglas production function describes the multiplicative relationship between a company’s total production (Y), technology (K), and labor (L). But we may need a new production function if companies can maintain significant amounts of production despite the labor term approaching zero. How will your organization produce in an increasingly autonomous future?