AI Isn't Your Employee. It's Your Performance-Enhancing Technology.

A fiberglass vaulting pole did not clear a world record. The athlete did. The pole just made a barrier the human body alone could not break reachable for the first time. That is the mental model AWS executive Ruba Borno wants executives to carry into their next AI budget meeting: AI is performance-enhancing technology, not a colleague, not a digital report, and not the next hire on the org chart.

Borno, who leads AWS's worldwide specialist organization, makes the case in a Fortune opinion column published Friday. The argument lands while executives are closing second-half 2026 headcount and AI budgets, and the stakes are concrete: the "AI as employee" language that has hardened over the last six months is now showing up in org charts, vendor decks, and board materials. Treating AI as headcount leads to misallocated budget, broken oversight, and org-chart fiction, Borno writes, and she is unfashionably direct about it.

The 1961 introduction of the fiberglass vaulting pole produced a run of world records in the decade that followed. Cycling aerodynamics reshaped the Tour de France. Graphite tennis rackets redrew the baseline game. Polyurethane swimsuits, banned in 2010, pushed records so far that some still stand. Borno's argument puts AI in the same category, and the workforce consequences follow from getting the bucket right: none of those tools replaced the athlete.

The reader reflex to discount Borno as a hyperscaler executive with skin in the AI game is reasonable. It is also the wrong one to stop at. The piece earns the reframe because the governance point survives a change of sponsor. If AI is headcount, a company can manage it the way it manages people: reviews, headcount plans, performance ladders. If AI is performance-enhancing technology, the right question becomes "what are the rules of the sport, and which of our people gets to break records inside them?" The first framing produces layoffs dressed up as automation. The second produces training, tooling, and process redesign.

Borno anchors the engineering argument in AWS's own numbers: a 1.65x year-over-year jump in shipping throughput, a 21% improvement in time to production, and AI assistance on more than 76% of code changes inside the AWS engineering org. Kiro, AWS's AI coding assistant, sits inside that workflow. A separate customer case, the World Surf League working with AllCloud, unified 50 years of fragmented performance data into a real-time AI platform and cut broadcast prep time by 80%. Borno frames both as evidence that the bottleneck is execution, not headcount, and titles the column's throughline "Move as Fast as Your Ideas."

Those engineering numbers are company-provided and not independently verified. A reader who treats them as audit-grade evidence will over-read the column; a reader who treats them as context for how AWS itself uses AI will use them correctly.

In December, Matt Garman said companies that replace junior staff with AI are pursuing "the dumbest idea I've ever heard", a stance Business Insider also reported on. Borno's column extends that line: the alternative to laying off junior staff is not hiring more humans to do AI's job, it is treating AI as a tool that lifts the humans already on the team. The two pieces share a vendor, a venue, and a stance that the "AI as colleague" frame is breaking the decisions that depend on it.

McKinsey's research on leading through AI disruption lands in the same neighborhood: the executives pulling ahead are the ones redesigning workflows, not replacing roles. The reusable rule survives attribution: when a workforce decision lands on a desk, ask whether the proposed change moves the pole or moves the athlete. If the answer is the pole, the line item belongs in tooling and training budgets. If the answer is the athlete, the company owes the human a different conversation.

Borno's column lands the same week H2 2026 plans finalize. The metaphor survives only as far as executives choose to apply it.