Voice Is the New Keyboard
Robinson Burkey@robinson
−1.7%
since Sep 30
The idea
I believe voice becomes the default way work gets entered into software over the next few years, with dictation and ambient transcription replacing typing across meetings, exam rooms, and factory floors. Adoption in big enterprises tends to lag the demos, so the basket leans on large platform names for ballast alongside the workflow software, chips, and pure-play voice AI companies that would benefit most directly.
What it owns
- The software where data entry happens today, from support calls to CRM fields to construction logs, stands to capture voice input as the primary way workers fill in information. · NICE LtdIt runs contact center software where conversations are the raw material, and its AI turns calls into records and actions without human typing, eliminating data entry friction at scale., ServiceNowIt runs the back-office workflows and tickets that get filled in by hand today, and its AI agents are adding voice input as the way workers complete those tasks., SalesforceIt is where white-collar data entry lives, and its Agentforce push lets salespeople talk updates into the CRM instead of typing them in by hand., TwilioIt is the plumbing developers use to add calls and voice AI to any app, so it gets paid whenever a new voice interface ships, regardless of who builds it., Procore TechnologiesIt is the construction software where field workers log inventory by talking into a clipboard app instead of typing, shifting voice from white-collar to blue-collar workflows., Five9It sells cloud contact centers where AI agents answer and log calls themselves, replacing manual data entry end to end as voice automation becomes the growth engine.28%
- The giants that own dictation engines, assistants, and the devices in everyone's pocket provide steady exposure to voice adoption across billions of users and devices. · MicrosoftIt owns Dragon, the dictation product already transcribing exam rooms and legal offices, plus Copilot voice across Office, giving it reach to hundreds of millions of workplace users if data entry shifts to talking., AlphabetIt has speech recognition models built into Android, Workspace, and its cloud, so any developer adding dictation to an app can license Google's speech technology as voice input spreads., Meta PlatformsIts Ray-Ban glasses are a consumer version of an ambient device that listens and acts on what it hears without a keyboard, making voice the primary input method., AppleIt already has microphones in hundreds of millions of ears through AirPods, and its on-device dictation reaches ordinary people at scale as voice input becomes primary.24%
- The microphones and low-power chips that every voice-first device needs to hear and process speech locally on the device will see rising value per unit as voice input spreads. · QualcommIt makes the chips that let phones, headsets, and glasses listen and run speech AI locally, so demand grows as ambient, always-listening devices become more common., Cirrus LogicIt makes the audio chips that capture and clean up sound in phones and earbuds, so microphones and audio processing matter more per device as voice input grows., AmbarellaIt builds chips that let devices understand what they see and hear at the edge, powering ambient systems that act on conversations in cameras, wearables, and industrial equipment., SynapticsIt sells the low-power chips that handle wake words and audio processing in earbuds, smart speakers, and connected devices, scaling with the number of listening gadgets.17%
- Understanding speech in real time requires serious AI compute, and these companies sell the silicon that powers speech models regardless of which application wins. · NVIDIAEvery dictation session, meeting transcript, and voice agent runs speech models on its GPUs, so inference demand grows with every spoken word in workplace software., Arm HoldingsIt designs the low-power processor cores inside every phone, earbud, and wearable that will do the listening, so higher-value chips ship as on-device speech AI spreads.11%
- Companies whose whole business centers on voice understanding offer the most direct and most volatile exposure to the shift from typing to talking. · SoundHound AIIt sells voice AI directly to businesses and cars, offering the most direct bet on enterprises adopting voice interfaces as the default way to enter work data and take orders., CerenceIt builds the voice assistants inside cars, one of the first places where talking beats typing because drivers have their hands on the wheel.10%
- Cash10%