Arrowhead AI 发布 Fletch 语音模型，支持 14 种语言并用于每日超百万通客服电话
Arrowhead AI Launches Fletch Voice Model for Multilingual Customer Calls
印度班加罗尔语音 AI 初创公司 Arrowhead AI 于 10 月 30 日发布自研 TTS 模型 Fletch，面向印度和东南亚的多语言客户通话，已部署在真实客服场景中，每天支撑超过一百万通电话。Fletch 支持英语、印地语、泰米尔语、泰卢固语、孟加拉语、马来语和印尼语等语言，可在同一句话内切换语言，并提供共情、致歉、平静、坚定等 12 种情感风格。公司称其延迟为 80 毫秒、不到 100 毫秒即开始生成语音，在品牌名、产品名、客户名和印度专有名词等选定词上发音准确率提升约 30%，目前用于催收、销售、续约和呼入客服，客户包括 Axis Bank、RBL Bank、IndusInd Bank、Tata 1mg、upGrad、KreditBee、Kissht 和 TVS Credit。
正文
Bengaluru-based voice AI startup Arrowhead AI on October 30 launched Fletch, an in-house text-to-speech (TTS) model designed for multilingual customer conversations across India and Southeast Asia. The model is already deployed across live customer calls, powering more than one million calls per day, the company said.
“Owning our voice layer means we can tune every millisecond and every emotion for how India and Southeast Asia actually speak,” said Devyani Gupta, Founder and CEO of Arrowhead AI.
“Fletch is our first step towards voice AI that sounds like it belongs here: 14 languages, and it starts speaking in under 100 milliseconds. We want every customer to hear a conversation, not a machine,” she added.
The company said that Fletch is designed to handle regional accents, code-mixed sentences and Indian proper nouns. According to Arrowhead AI, Fletch supports languages including English, Hindi, Tamil, Telugu, Bengali, Malay and Indonesian, and can switch between languages within a sentence. The company added that the model also adapts its vocal tone to the context of a conversation, with 12 emotion styles, including empathetic, apologetic, calm and firm.
The company claims that the model has been trained on thousands of hours of curated regional speech and recordings from professional voice artists. Arrowhead said Fletch operates at a reported latency of 80 milliseconds and begins generating speech in under 100 milliseconds.
Arrowhead AI said Fletch is deployed across customer interactions involving collections, sales, renewals and inbound customer service. The company works with enterprises including Axis Bank, RBL Bank, IndusInd Bank, Tata 1mg, upGrad, KreditBee, Kissht and TVS Credit.
According to Arrowhead, Fletch delivers approximately 30% better pronunciation accuracy on selected terms, including brand names, product names, customer names and Indian proper nouns. The company said the improvement can reduce instances in which customers ask voice agents to repeat themselves.
Vengadanathan Srinivasan, Co-founder and CTO of Arrowhead AI, said, “We built on the latest research in speech generation and trained it on thousands of hours of regional speech, with real voice artists. The result runs at 80 milliseconds of latency and adjusts its tone to what’s being said.”
Fletch is built on a DiTAR-based architecture adapted and trained by Arrowhead. The model operates alongside the company’s speech recognition system, language model and dedicated address model within its voice AI platform.
Arrowhead is expanding Fletch’s capabilities across Southeast Asia, including support for Malay and Indonesian. The company plans to demonstrate the model at Singapore FinTech Festival 2026, scheduled for November 18–20 at Singapore EXPO.