Álvaro de Nicolás
← All essays

Executive memo · AI & operating models

Generative AI + 'Brain Fry'

When AI intensifies work instead of simplifying it · Álvaro de Nicolás · 2026

Generative AI + Brain Fry visual summary

The finding

A study of 1,488 full-time US workers shows the same tool that should reduce cognitive load is, in many deployments, increasing it. The mechanism is oversight: when people are asked to supervise too many AI tools, agents and outputs, AI stops being a productivity lever and becomes a source of mental fatigue.

AI reduces burnout when it removes repetitive work. It creates burnout when people are forced to monitor too many tools, agents and results.

What the data shows

The business cost

What helps vs. what hurts

Helps: AI replacing repetitive work, a clear AI strategy, training and support, managers answering questions, teams integrating AI into real workflows, work-life balance.

Hurts: too many agents to supervise, pressure to use AI for everything, more workload because of AI, unclear expectations, "figure it out yourself" management, activity-based metrics instead of impact-based ones.

Five actions for leaders

  1. Cap AI oversight load. Do not stack endless agents on a single person.
  2. Set explicit expectations on AI use and workload.
  3. Measure impact, not tokens, clicks or lines of code.
  4. Redesign workflows for human + AI together — not human supervising AI.
  5. Train people in framing, prioritisation and decision-making — the parts AI cannot do.

Bottom line

AI is a productivity multiplier up to a point — and a cognitive tax beyond it. The companies that win will not be those deploying the most tools, but those designing the supervision load deliberately.