People still treat "sounds human" as a quality metric. But that metric hasn't matched reality for a long time.
Today, what actually matters is whether users instantly understand the message, whether the product gains traction, and whether support costs actually drop in the target market.
If "natural" achieves that — perfect. If it doesn't, then it's not craftsmanship. It's vanity.
Translation is about effectiveness
You often can't tell if a document was translated by a human or by AI — but who cares, if users still can't act on it or move any metric?
If users understand it instantly and adoption grows, does it really need to "sound human"?
That's the shift everyone feels but few want to say out loud.
Some categories are different
Of course, some industries are a different game entirely — beauty, luxury, hospitality. Their content isn't meant to inform; it's meant to make users dream. Those categories genuinely need a crafted emotional tone.
SaaS doesn't.
AI is already "natural enough" for a huge portion of content. But it still can't decide whether an output is good enough for the goal, technically fine but strategically wrong, or not acceptable and requiring a rewrite.
Judgment, not fluency
That judgment — not fluency — is where the actual value is.
Translation quality has a new definition. The question is no longer "does this sound natural?" It's "does this work?"