Factory work has never been exactly glamorous. But it can be quite lucrative, so long as you correctly balance your time and efforts like a gymnast perched on a narrow beam.

To appreciate the dangers of that balancing act, imagine this frustrating scenario. Every Monday, as head of an industrial plant, you watch with dismay as your top people disappear into a pile of Requests for Quotes (RFQs), knowing full well that all their hard work may result in nothing.

“The problem is simple,” explains Daryl Edwards in our interview. “Factories waste time figuring out a price for projects they might not even win.”

Edwards understands this dilemma. Before founding Toronto-based Agent Impact, a manufacturing AI firm, he was a plant manager as well as VP of operations, helping to scale Peru’s largest shoe exporter.

Though the formidable challenge Edwards names is real, so is the possible solution now emerging: Agentic AI. To appreciate the fix, we must first go back in time and across the world. Following World War II, Japan was committed to rebuilding its economy after so much devastation.

The defeated nation was short on resources and pinched for space to process inventory. It couldn’t afford to pursue America’s more wasteful production model. The States’ industrial approach often ran on buffer stock and big inventories. Flush with a booming economy, American warehouses could afford to sit full against unpredictable orders and stalled wait times for parts.

Japan didn’t have that luxury. So it developed the Toyota Production System (TPS), built on a principle it called Just-In-Time production: “making only what is needed, when it is needed, and in the amount needed,” according to the Toyota Motor Corporation .

The idea spread like wildfire.

“After Toyota's outstanding success with TPS in meeting and exceeding American manufacturing production at significantly lower costs, manufacturers around the globe rapidly adopted and emulated TPS-based practices throughout the 1980s, 1990s, and beyond,” explains EBSCO . It cites a 2006 survey of U.S. manufacturers that found “Just-In-Time supplier deliveries were at the top of the list of most commonly used methods for managing inventories.”

Revelatory as this model is to global manufacturing, it still suffers from one major limitation in 2026: the bottleneck caused by today’s quoting system.

To see why, picture a customer that needs 10,000 ball bearings. Yesterday . Naturally, they contact several factories for a quote. The one that responds fastest with a fair price will win the job. Simple, right? Nope. Determining such prices isn’t easy. Before a factory can commit to a price, seasoned workers must half-design the part to determine what it’ll cost.

The brutal part is they have to do this on a contingency basis. Work too slow, too methodically, and your competitor will beat you to the punch. Shortcut the process and you may end up with a faulty quote you can’t fulfill, wrecking your profitability and reputation.

This brings us back to the wonders of Agentic AI. Edwards recently pioneered an approach he hopes will represent factory’s newest reinvention. The process requires breaking quoting down to what he dubs “atomic steps,” the tiniest possible units of work.

“Quoting comes down to 15 or 20 of them,” he says. “They can include everything from confirming an email request contains the pertinent info to extracting machine and labor requirements.” Providing an added layer of assistance to factory workers, Agentic AI can perform other valuable actions, like assembling all the items a CAD designer needs before they start drawing.

Quoting falls within just one of several categories where this atomic approach applies. Others include planning and scheduling, quality, engineering and design, production visibility, and finance, HR, and admin. Regardless of the specific use case, agents fill the gaps between what a team knows and what it never seems to have sufficient time to discover.

The result is a dynamic new way of working that saves time and resources, enabling factories to do more with less, a key component of the Micro Enterprise Model I recently wrote about for Forbes . “We’re not replacing people,” Edwards is quick to point out. “This is about lifting the dull, information-heavy tasks off the bottom of the stack of people’s time, so they move up and do the higher-value stuff, including relationships and strategy.”

This distinction points to something bigger. What's being reimagined is how any company built on Agentic AI may design future workflows, a subject Peter Diamandis took up recently on his podcast, Moonshots , in a segment called “The New Era of Jobs: Organizational Singularity.”

Edwards calls the idea behind this the Agentic Ladder , a simple way to answer what people mean by an AI agent. As he told me, “each rung adds a real technical capability, from a passive reference up to a self-directing colleague. Each does more of the work for you while needing less oversight.”

What follows maps the rungs to employee seniority, with the technical leap listed parenthetically:

-L0, Reference (lookup only): a book on the shelf. It retrieves stored information on request, with no generation and no action. -L1, Collaborator (most people's ChatGPT today): a new hire in week one. It generates text, code, and images on request, one turn at a time. You review every output. -L2, Consultant (chains tool calls): a junior employee. It strings several tool calls together to finish a routine task and escalates the edge cases to you. -L3, Approver (what "agent” usually means): a senior employee. It triggers on events, uses tools, writes back to your systems, and summarizes work for your sign-off. -L4, Observer (runs continuously): a department manager. It runs the function on its own against goals and guardrails. You review the KPIs rather than individual actions. -L5, Investor (self-improving): a self-developing leader. It watches outcomes and rewrites its own playbooks, improving without being reprogrammed.

This hierarchy has implications for corporate America beyond the manufacturing space it was formulated in. We can imagine a professional services firm like a marketing agency using it to execute workflows. The bigger takeaway is that Agentic AI is transforming how future companies operate, paving the way for increasingly productive human/machine hybrid work dynamics, no matter the sector.

“There’s a metaphor I keep coming back to that helps explain the future of work,” says Edwards. “With Agentic AI, you are not buying a tool. You are hiring a teammate and promoting them over time.” The reason many factories, even modern ones, remain stuck, whether in quoting, scheduling, or quality, is that they stop at L1. At best. “The real payoff arrives as you move work up the ladder, where the agent does the task for you instead of just helping you do it faster.”

Once upon a time, America rose to global prominence as the leading manufacturing light. Even before Just-In-Time production, the United States led the way as a model of innovation, shaping the way the world works. Now that Agentic AI is here, we have the opportunity to provide the next big leap in manufacturing and more, a promising development that may boost productivity gains and usher in a rise in standard of living we are now beginning to grasp.

Yes, factory work may never be glamorous, but thanks to Agentic AI, tomorrow’s companies can seize this exciting technology to win back time, that rarest of luxuries on any plant floor.