In 2022, she was hired to build AI operations at a health-tech startup. They were pioneering the use of AI in healthcare, and her role was to oversee the human side of the transition — managing the workflows, the quality control, the gradual integration of machine assistance.

Then GPT-4 launched.

Within a short period of time, she realized her role no longer made sense. The tasks she had been hired to coordinate and supervise — reviewing outputs, managing quality, bridging the gap between AI systems and human workflows — were being absorbed by the models themselves.

Her employer came to the same conclusion. There was no plan to retrain her. No conversation about redeploying her skills into a new version of the work. Her job simply disappeared.

"I was a cost that disappeared," she says. "The underlying capability question — what should this work become? — was never asked."

She says this not as a warning but as context. When she looks at the wave of mass layoffs being justified as AI transformation, she's not reading about it from a distance. She's been on the other side of that decision.

What she understands now that she didn't fully see then is that her employer wasn't transforming — they were optimizing. Layoffs offer clean math: immediate cost savings, a simple story for boards eager to see returns on AI investments. What they don't deliver is increased capacity, creative leverage, or new kinds of work.

Today she heads up AI Operations at Pearl, an AI company for independent professionals, where they've taken a different path: upskilling employees, reshaping roles, and having uncomfortable conversations earlier than most companies are willing to.

She works closely with a technical writer who once asked: "AI can do a lot of my work for me — so what is my job now?" Instead of avoiding the question, they answered it together. The writer now operates like an entire technical writing department, using AI agents for proofreading, editing, and standardization. She also owns the company intranet — keeping it alive with AI-powered updates, a function that usually fails because it depends on constant manual work.

"The companies that are truly transforming with AI are doing something far more difficult than issuing headcount reductions," she says. "They're acknowledging that work itself is changing and actively designing for it. The rest? They're just laying off their way to transformation because it's easier than rewiring how work gets done."

She knows where that road leads. She's been there.