Just saw this metric published in the 2025 corporate highlights from BlueFocus (a massive global marketing and PR conglomerate). They explicitly stated that their internal AI systems have now "surpassed human performance in 85% of marketing operational scenarios."
What strikes me here is the disconnect in how we talk about AI. In the AI community, we spend endless hours arguing over MMLU scores, ARC challenges, or SWE-benchmarks. But out in the enterprise world, they aren't testing AI on abstract math—they are measuring it against actual human KPIs in day-to-day white-collar operations (media buying, copywriting, SEO optimization, and basic asset generation).
When a $10B enterprise publicly states to its stakeholders that AI is beating their human baseline 85% of the time in practical applications, it feels like the tipping point for white-collar automation has already happened behind closed doors. The technology is capable now. We are simply waiting for the macroeconomic data (and the job market) to catch up to reality.
For those working in tech or enterprise sectors, are you seeing similar internal metrics in other industries? Is this 85% number just corporate hyperbole, or is the "silent automation" of white-collar work happening much faster than the general public realizes?