Indeed Engineers Write 50% More Code With AI. Who Gets the Hours?

By Daniel Okafor

At the Fortune Most Powerful Women conference in October 2026, Indeed chief revenue officer Maggie Hulce reported a concrete number: software engineers at her company write 50% more code with AI. The same tools save Indeed employees six hours per week.

Six hours is roughly 15% of a standard 40-hour workweek.

Now executives from PepsiCo, Dropbox, and Bristol Myers Squibb ask a harder question: who keeps them?

What does the productivity boom mean for workers?

PepsiCo's Athina Kanioura pushed back against measuring AI through financial returns alone. She holds two titles: global chief strategy and transformation officer and CEO of PepsiCo Latin America. She described the relentless pursuit of ROI on every AI project as "death by a thousand cuts." A rigid business case, she argued, overlooks benefits that don't fit neatly on a quarterly earnings slide.

Each new AI application is judged on near-term dollar impact, so projects with benefits that are harder to quantify — creativity, employee well-being, brain health — often get cut at the proposal stage.

Hulce framed the upside in plain language. "We're moving faster, we can innovate faster, we can invent faster, we can do fun work more and more," she said. Indeed is steering the saved hours toward creative work and the projects employees say they enjoy most.

Why are some leaders worried about cognitive decline?

U.S. employees who frequently use AI report higher burnout than those who don't, according to 2026 data from Quantum Workplace. The efficiency surge carries a hidden cost.

Dropbox chief people officer Melanie Rosenwasser said the $8 billion cloud company studies the risk of "cognitive atrophy" as workers outsource their thinking to AI. The HR team has introduced memory exercises, attention-span training, and puzzles for staff. Dropbox also experimented with device-free "analog" off-sites and brought in a breathwork specialist for its executive team.

The pattern she described is straightforward but uncomfortable: AI handles routine reasoning, so workers practice reasoning less, and the cognitive skills they built through years of work begin to erode from lack of practice.

"When you outsource your thinking to AI, you become less intelligent in the process," Rosenwasser said, citing research from MIT professor Eric So. She added: "So we need to be very deliberate about how we're using these tools."

Who keeps the productivity gains?

Bristol Myers Squibb EVP and chief people officer Amanda Poole is weighing how to split the productivity dividend between company and workforce. The drugmaker tests a "well-being quotient," a framework for deciding what portion of automated gains should support employees and what portion stays as profit.

The math is straightforward on paper but harder in practice: a 20% productivity gain at a single worker can be split many ways, and most firms default to keeping most of it for shareholders rather than staff.

"Let's say we're going to get 20% productivity [with AI]: What part are we going to take back as a company, and what part are we going to give back?" Poole said. She added: "We haven't solved it, so more to come."

What changes next?

Four companies run separate tests on the same underlying problem:

• Indeed reinvests saved hours into creative work.
• Dropbox guards against cognitive decline through training programs.
• Bristol Myers designs an internal metric to govern the split.
• PepsiCo resists tying AI projects to short-term financial returns at all.

Each approach tests the same hypothesis in a different shape. The question for the rest of 2026: can AI's productivity dividend flow to workers without hollowing out the skills or well-being that make the work worth doing in the first place?