Voice-to-text tools have been around for decades, but most office workers still rely heavily on keyboards. Despite the technology existing for a long time, many find voice dictation less accurate or feel awkward talking to AI assistants at work. Wispr, a startup with significant funding and a high valuation, aims to change that. Its new Flow tool turns speech into text across desktop and mobile apps, including Slack, Gmail, and AI chatbots like ChatGPT. It automatically cleans up transcripts by removing filler words and fixing typos, promising better accuracy than existing dictation options.

How Wispr’s Technology Stands Out

Wispr claims its voice-first models are trained on how people really speak, which helps improve transcription quality. Instead of adding speech recognition as an afterthought to general language models, Wispr develops models that understand transcription, formatting, and user intent all at once. For editing, it uses an open-source model called Llama 3.1 combined with proprietary AI from companies like OpenAI. This setup allows Flow to accurately track speech at low volumes, potentially making it easier to talk in open-plan offices without disturbing colleagues.

CEO Tanay Kothari notes that Flow can pick up quieter speech, making it more practical for everyday office use. The tool is gaining popularity, with about 125 new customers each week, spanning industries from legal teams to sales and engineering. Companies are adopting it to boost productivity, especially as more workers incorporate voice commands into their workflows.

The Impact on Workplace Productivity

Many users report significant changes after starting to use Flow. On average, they cut their daily typing from five hours down to three. Over five months, 72% of users’ computer activity shifts to using Flow, with less reliance on traditional keyboards. This shift suggests that voice tools might become a core part of how people work, not just a niche feature.

AI tools serve as an entry point for many users. Once they try using Flow with chatbots like ChatGPT or Cursor, they often realize they want to use it everywhere. Soon, they’re employing it for Slack messages, emails, and document drafting. This widespread adoption could change how workplaces handle communication and documentation in the future.

Use Cases and Industry Adoption

Flow is already being used in diverse settings. For example, at Yahoo, the chief legal officer uses it to draft legal documents and comment on Google Docs and Word files. At Ramp, the analytics team employs Flow across engineering workflows and documentation tasks. Such examples show how versatile the tool can be across roles and industries.

As voice technology continues to improve, it could become a standard part of daily work routines. The ability to speak naturally and have AI handle transcription and editing may make office work more efficient and less frustrating. Wispr’s progress suggests that a post-keyboard office might not be so far off after all.

Inspired by

• https://www.computerworld.com/article/4107331/wispr-ceo-interview-post-keyboard-office.html

Sources

• pexels.com
• wisprflow.ai
• techcrunch.com
• wisprflow.ai
• wikipedia.org