Arrowhead AI Launches Fletch Voice Model Powering Over 1 Million Customer Calls Daily

Arrowhead AI Deploys Fletch Voice Model Across Live Customer Operations

Bengaluru-based voice AI startup Arrowhead AI launched Fletch on October 30, an in-house text-to-speech model engineered for multilingual customer conversations across India and Southeast Asia1

1

2

2

. The model is already deployed across live customer calls, powering more than one million calls per day1

1

. Fletch represents a shift toward regionally-tuned conversational AI that addresses the linguistic complexity of customer service automation in diverse markets.

Technical Architecture Delivers 80ms Latency Performance

Fletch operates on a DiTAR-based architecture adapted and trained by Arrowhead AI1

1

2

2

. The model achieves 80ms latency and begins generating speech in under 100 milliseconds1

1

2

2

. This performance threshold matters for real-time customer interactions where delays disrupt conversational flow. Co-founder and CTO Vengadanathan Srinivasan explained that Fletch was trained on thousands of hours of curated regional speech and recordings by professional voice artists, using research in speech generation2

2

. The model operates alongside the company's speech recognition system, language model and dedicated address model within its voice AI platform1

1

.

Handling Code-Mixed Sentences and Regional Linguistic Patterns

Fletch supports 14 languages including English, Hindi, Tamil, Telugu, Bengali, Malay and Indonesian, and can switch between languages within a single sentence1

1

2

2

. The text-to-speech model is designed to handle regional accents, code-mixed sentences and Indian proper nouns1

1

2

2

. Arrowhead AI claims Fletch delivers approximately 30% better pronunciation accuracy on selected terms, including brand names, product names, customer names and Indian proper nouns1

1

2

2

. This improvement can reduce instances where customers ask voice agents to repeat themselves1

1

. The model also adapts its vocal tone to conversational context with 12 emotion styles, including empathetic, apologetic, calm and firm1

1

2

2

.

Enterprise Deployment Across Banking and E-Commerce

Fletch is deployed across customer interactions involving collections, sales, renewals and inbound customer service1

1

2

2

. Arrowhead AI works with enterprises including Axis Bank, RBL Bank, IndusInd Bank, Tata 1mg, upGrad, KreditBee, Kissht and TVS Credit1

1

2

2

. Founder and CEO Devyani Gupta stated, "Owning our voice layer means we can tune every millisecond and every emotion for how India and Southeast Asia actually speak. We want every customer to hear a conversation, not a machine"1

1

2

2

.

Expansion Plans Target Southeast Asian Markets

Arrowhead AI is expanding Fletch's capabilities across Southeast Asia, including enhanced support for Malay and Indonesian1

1

2

2

. The company plans to demonstrate the Fletch voice model at Singapore FinTech Festival 2026, scheduled for November 18-20 at Singapore EXPO1

1

2

2

. Watch for how financial institutions and e-commerce platforms in Southeast Asia adopt this technology, particularly as regional competition in voice AI intensifies. The focus on regional linguistic nuances positions Arrowhead AI to capture market share where global models struggle with local context.