For most of business history, output scaled with headcount. More people meant more work done. AI is breaking that equation — and two business owners documented exactly what the new math looks like.
One counted every dollar. One counted the people who were left standing.
Trave Harmon, Triton Technologies: 38 Projects. $563. $767,983 Back.
Trave Harmon is the CEO of Triton Technologies, a managed IT and cybersecurity company serving small and mid-sized businesses in the Northeast. Between February 25 and April 16, 2026 — 50 calendar days — he used Claude as his primary development partner to complete 38 operational projects across every function of the business. Alone.
The numbers he published are unusual enough that he included the verification sources.
Total AI spend: $563.30, confirmed via card records. First-year value returned: $767,983, across two categories:
$364,858 in annual recurring savings. Software subscriptions cancelled and replaced with custom-built alternatives. A Zapier dependency his company had maintained for eight years: gone. CoSchedule, Buffer, and Hootsuite — three separate tools for social scheduling — replaced by one system. An outsourced bookkeeper replaced by an AI-assisted process his team now runs internally.
$403,125 in one-time value created. The largest single project: a complete rewrite of Triton’s 393-page website. A professional web agency would have priced that engagement between $120,000 and $250,000 based on published per-page rates. Harmon did it in one week with AI. Page load time dropped from 7,000 milliseconds. Hundreds of SEO violations fixed.
He also caught a Massachusetts sales tax overpayment — $13,043 per year, verified against ACH and tax records — that had been sitting there unnoticed. And a workers’ comp reclassification worth another $1,075 annually, confirmed against 140+ payroll emails.
The ROI works out to 136,338%. Harmon published the math specifically because the number sounds implausible. The documentation is there for anyone who wants to check it.
Takeaway: He didn’t replace anyone. He redirected what one motivated person with the right AI partner could do in less than two months.
Eric Vaughan, IgniteTech: The 80% Who Didn’t Come With Him
Eric Vaughan runs IgniteTech, an enterprise software company. In early 2023, he looked at his global workforce and saw a problem — not a performance problem, a velocity problem. He believed AI was an existential transformation and that failing to move fast would doom the business.
He called an all-hands meeting. His message was direct: everything would now revolve around AI. The company would reimburse for AI tools, prompt-engineering courses, and outside experts. He wasn’t asking. He was announcing.
What followed was harder than he expected. “Changing minds was harder than adding skills,” he told Fortune. Over the course of 2023 and into early 2024, IgniteTech replaced nearly 80% of its staff — hundreds of people — not because their roles disappeared, but because they wouldn’t make the shift.
Two years later, in a 2025 interview, Vaughan said he’d do it again.
The 20% who remained didn’t just survive. They inherited the full operational capacity that used to require five times the headcount.
Takeaway: Vaughan’s story is the industrial version of what Harmon did in a spreadsheet. The people who adapted didn’t just keep their jobs — they became the department.
Both of these happened in the last two years. Both are documented. Both point at the same pattern: the ratio of people to output is shifting, and it’s shifting fast. Harmon proved it costs less than $600 to run a company for 50 days at scale. Vaughan proved that when 20% of a workforce commits fully, the other 80% becomes optional.
The question for every business owner isn’t whether this is happening. It’s where they are in the ratio right now.