A Headset Company Says the Keyboard Is Dying. Here's the Catch.

Jabra just launched a new headset line and, along with it, a bold prediction: voice AI will become the default way we work by 2028. No more typing emails, drafting documents, or prompting AI tools. Just talking.

It's a compelling story. It's also worth noticing who's telling it.

The headline stat comes from a joint study with the London School of Economics, and it's genuinely interesting: 14% of participants already prefer voice over typing when using GenAI tools. That's a real, measurable shift. It's also a long way from "the keyboard is a Victorian construct," which is how Jabra's own product marketing manager, Richard Trestain, put it. He's not exaggerating on the history, the modern typewriter really does trace back to 1867. Whether that makes today's keyboard obsolete or just old is a different question entirely.

Trestain is a genuine power user of this shift. He runs his entire workday through a foot pedal and a voice AI dictation app called Wispr Flow, paired with Jabra's own headset. By his own admission, he's an early adopter, not a representative sample of the average knowledge worker.

The actual engineering story here is more credible than the "keyboard is dead" framing. Jabra's ClearVoice technology, a deep neural network trained on 60 million sentence examples across noisy environments, claims to capture voice input with over 99% accuracy. Crucially, it's not learning your specific voice like a personal assistant would. It's trained on the acoustic structure of human speech in general, then baked directly into the headset's chip. That's a meaningfully different (and more privacy-friendly) approach than most voice AI products on the market, and it's tested using calibrated mannequins in anechoic chambers designed to simulate real-world noise, which is a genuinely rigorous way to validate a hardware claim.

The three new models, the Evolve3 45, Evolve3 65, and Evolve3 65 Flex, aren't one-size-fits-all. Jabra's president Calum MacDougall makes a fair point: desk-based workers, open-office employees, and hybrid commuters have fundamentally different audio needs, and treating them identically has always been a design compromise, not a solution.

Here's the honest read. Voice-driven interaction with AI tools is clearly growing, and the hardware innovation solving for noisy, imperfect real-world audio capture is legitimate progress. But "keyboard-less by 2028" is a prediction from a company that sells headsets, made using a study that shows 14% adoption, not a majority. Worth watching the trend. Worth being skeptical of the timeline.

Do you see yourself talking to your computer instead of typing within the next few years, or is this further off than the pitch suggests?