内容来源：https://aibusiness.com/generative-ai/speech-to-text-ai-specialist-valued-at-2b

内容总结：

美国人工智能初创公司Wispr近日宣布完成2.8亿美元B轮融资，公司估值达到20亿美元。这家总部位于旧金山的企业主营AI语音输入平台，其核心产品Wispr Flow允许用户通过语音而非文字与移动及桌面设备交互，自去年5月完成3000万美元A轮融资后，市场热度持续攀升。

该公司联合创始人兼首席执行官塔纳伊·科塔里表示，一年多前市场对语音技术普遍持怀疑态度，因为用户过去15年尝试各类语音工具均未获得理想体验，而如今Wispr Flow凭借高质量输出显著节省用户时间，平台不仅实现基础转写，还能格式化用户思路、学习个性化词汇并模仿用户说话风格，吸引约60%的非英语用户使用。

Wispr计划推出全新专有语音模型Canto，以应对嘈杂背景、音乐播放或浓重口音等复杂语音场景。同时，公司近期发布会议记录工具Notetaker，进一步拓展应用场景。本轮融资由长期投资者Menlo Ventures领投，现有投资方Notable Capital、NEA、Neo Ventures等参投，新增投资者包括Acrew Capital、Forerunner Ventures等。达拉斯牛仔队四分卫达克·普雷斯科特、匹兹堡钢人队接球手DK·梅特卡夫及演员奥利维亚·邓恩等文体明星亦参与投资。新资金将主要用于产品研发及提升识别准确率。

中文翻译：

由谷歌云赞助

选择你的首批生成式AI用例

要开始使用生成式AI，首先应聚焦于那些能够改善人类获取信息体验的领域。

这家供应商正属于吸引大量融资的本土AI初创企业浪潮中的一员。

美国初创公司Wispr已完成一轮2.8亿美元的融资，估值达20亿美元，其AI驱动的听写平台需求持续快速增长。

这家总部位于旧金山的供应商周一在其网站上发表了一篇帖子，确认了这一B轮投资。此前，该公司在去年5月完成了3000万美元的A轮融资。

该公司的核心产品是听写应用Wispr Flow，它能让用户通过语音而非文字与移动端和桌面端环境进行交互。然而，Wispr Flow之所以在同类工具中脱颖而出，在于其受欢迎程度。

在帖子中，联合创始人兼首席执行官塔奈·科塔里表示，就在一年多前，人们对语音技术普遍持怀疑态度。他说：“这种怀疑是很自然的。人们尝试语音工具已经15年了，每次得到的结果都一样，所以合理的假设是这次也会让他们失望。”

据该供应商称，用户现在表示Wispr Flow为他们节省了大量时间。

Wispr将此归因于输出质量，该平台不仅仅局限于基础转写，还能格式化用户的思路、学习用户的词汇，并生成符合用户说话方式的文本。全球覆盖也加速了其普及速度，该应用约60%的输出为非英语内容。

公司还计划进行改进，包括正在开发的一款新的专有语音模型——Canto——该模型能够应对嘈杂环境、播放音乐或带有浓重地方口音等困难条件下的语音识别。

在Flow专注于用户对着电脑说话的语音识别的同时，Wispr最近还发布了Notetaker，这是一款用于与他人开会的工具。

Wispr表示，将把最新一轮融资用于进一步的产品开发，并持续提升准确性。领投B轮的是长期投资者Menlo Ventures，回归投资者包括Notable Capital、NEA、Neo Ventures、8VC和MPV Ventures。

新投资者包括Acrew Capital、Forerunner Ventures、Together Fund和Goodwater Capital。其他投资者还包括多位名人和体育明星，如达拉斯牛仔队的达克·普雷斯科特、匹兹堡钢人队的DK·梅特卡夫，以及电视剧《海滩救护队》中的体操运动员利维·邓恩。

英文来源：

Sponsored by Google Cloud
Choosing Your First Generative AI Use Cases
To get started with generative AI, first focus on areas that can improve human experiences with information.
The vendor is part of a wave of native-AI startups that are attracting strong funding.
U.S. startup Wispr has completed a $280 million funding round at a $2 billion valuation as interest in its AI-powered dictation platform continues to grow rapidly.
The San Francisco-based vendor confirmed the Series B investment in a post on its website on Monday. It follows a $30 million Series A round last May.
The company’s core product is the dictation app Wispr Flow, which enables users to interact with mobile and desktop environments using speech rather than text. What makes Wispr Flow stand out, though, compared to similar tools, is its popularity.
In the post, co-founder and CEO Tanay Kothari said that just over a year ago, there was widespread skepticism about voice tech. He said: “That skepticism was natural. People had been trying voice tools for 15 years and getting the same result every time, so the reasonable assumption was that this one would disappoint them too.”
Users now say Wispr Flow saves them significant time, according to the vendor.
Wispr attributes this to the quality of the output, with the platform going beyond basic transcription to format users’ thoughts, learn their vocabulary and produce text that mirrors their way of talking. A global reach has also accelerated uptake, with about 60% of the app’s output non-English.
Improvements are planned, including a new proprietary speech model -- Canto -- being developed that can cope with voices in difficult conditions such as background noise, music playing or when strong local accents are used.
And while Flow concentrates on the speech users direct at their computer, Wispr also recently released Notetaker, a tool intended for use in meetings with other people.
Wspr said it will dedicate the latest funding toward further product development and continued accuracy improvements. Leading the Series B round was longtime investor Menlo Ventures, with returnees including Notable Capital, NEA, Neo Ventures, 8VC and MPV Ventures.
New investors include Acrew Capital, Forerunner Ventures, the Together Fund and Goodwater Capital. Other investors include several celebrities and sports stars, including the Dallas Cowboys' Dak Prescott, the Pittsburgh Steelers' DK Metcalf, and TV show Baywatch’s gymnast Livvy Dunne.