Garrett stepped into the shop and tasted the air before anyone spoke. It was sour with coolant and the faint tang of metal dust, but worse than that was the silence. Machines stood idle. A dozen men lingered near the break room door, arms crossed, their voices low. The company was dying, and they knew it.
He’d seen this scene before, in other towns, other shops. A place where paperwork moved slower than steel, where meetings produced words and nothing else, where no one could say exactly what they were working toward.
The plant manager, Dan, came forward, wiping his palms on a stained button-down. “You’re Garrett, right? They tell me you’re supposed to… help.”
“That’s the idea,” Garrett said. He didn’t smile. The situation didn’t call for it. “Let’s start with what’s not working.”
The first meeting was a mess. Everyone spoke over one another. Complaints piled on: overdue orders, confused paperwork, scheduling errors, emails lost in inboxes. Garrett typed quietly, barely raising his eyes.
After an hour, Dan asked, “So? You seeing what I’m seeing?”
“I’m seeing a lot,” Garrett replied. He tilted his laptop so the room could see the glowing screen. The AI-generated transcript scrolled in real time, speaker by speaker. Every action item highlighted in yellow. Every unresolved issue pinned to the side.
The room went silent.
“I don’t need anyone to repeat themselves,” Garrett said. “You’ll get a copy of this before you’re back at your desks. No excuses.”
One machinist muttered, “Hell, that’s better than Linda’s notes,” and drew a laugh from the others. Linda, the HR manager, blushed but didn’t disagree.
By the next morning, Garrett had sent out a consolidated task list with deadlines. Not his words, not exactly. The system drafted them overnight, and he nudged them into shape with half an hour of editing. When Dan saw it, he shook his head.
“You did all this already?”
“Not me,” Garrett said. “The computer.”
Two weeks later, things looked different. Machines hummed again. Schedules lined up. Nobody waited around for instructions because every morning, their inboxes carried a digest of priorities written in plain English. Shop floor supervisors stopped arguing in meetings because the AI captured agreements in real time, turning them into protocols by the next day.
“What’s this called, what you’re doing?” Dan asked one afternoon.
“Doing the AI,” Garrett said.
“That’s not a thing.”
“It should be.”
The turnaround was fast enough to get noticed. A defense contractor came calling, looking for suppliers who could spin up parts with precision and speed. Most shops in the region were in shambles, drowning in backlogs. But this one had cleaned house, documented every process, and could show efficiency metrics on command.
Garrett sat in the back of the conference room during the pitch, listening. When the contractor rep asked how quickly they could pivot production to new specifications, Dan glanced at Garrett.
“We’ve already run the simulations,” Garrett said, sliding a printout across the table. “Here’s the timeline, risk assessment, and QA process. Drafted yesterday.”
The contract landed by the end of the week.
The shop changed after that. New machines arrived. Younger workers, curious and ambitious, joined the crew. Old-timers shook their heads at how fast things moved, but no one complained about steady paychecks.
One evening, Dan found Garrett still at his desk, laptop open, the AI humming through another pile of protocols.
“You sticking around long-term?” Dan asked.
Garrett closed the lid. “Not my job. I get places pointed in the right direction, then I leave.”
Dan frowned. “And what happens when the AI breaks?”
Garrett smiled this time, just faintly. “Then you learn to do the AI yourself.”