# Stretto's AI Voice Agent Tara Boosts Lead Conversion and Reduces Costs

> Discover how Stretto's AI voice agent, Tara, utilizes Amazon Nova 2 Sonic to transform bankruptcy intake by engaging consumers in meaningful conversations, thereby gathering critical financial information while ensuring privacy and empathy.

**Source**: aws.amazon.com | **Published**: 2026-11-18 | **Type**: case_study

## Key Facts

- Stretto achieved a 20% increase in lead conversion rates, enhancing competitive positioning in bankruptcy services.
- A 95% reduction in voice costs allows Stretto to improve margins, indicating strong financial performance.
- The AI agent's 96% data extraction accuracy reveals operational efficiency, reducing human error in lead qualification.
- Natural conversation flow and emotional adaptability provide a competitive advantage in sensitive client interactions.
- Phased deployment strategy mitigates risk, suggesting a strategic shift towards data-driven decision-making in AI projects.

## Summary

\## Summary
Stretto, a company in the bankruptcy services industry, faced challenges with low-quality leads generated from a web form on its Bankruptcy.com site. To address this, they developed Tara, an AI voice agent using Amazon Nova 2 Sonic, which conducts detailed bankruptcy intake interviews over the phone. This solution resulted in a 20% increase in lead conversion rates and a 95% reduction in voice costs.

\## Background
Stretto operates in the bankruptcy services sector, primarily through its Bankruptcy.com platform. Before deploying the AI voice agent, the company relied on a basic web form that collected limited information from consumers considering bankruptcy. This method produced leads with thin margins, as bankruptcy attorneys preferred more comprehensive data to assess potential clients.

\## Challenge
The primary challenge was the significant gap between the basic information collected via the web form and the detailed financial data required by bankruptcy attorneys. Stretto needed a solution that could gather sensitive financial details through a conversational approach, as longer web forms led to high abandonment rates. Staffing a human intake team was also cost-prohibitive.

\## Solution
Stretto collaborated with the AWS Generative AI Innovation Center and CI&T to create Tara, an AI voice agent powered by Amazon Nova 2 Sonic. Tara conducts bankruptcy intake interviews over the phone, guiding callers through a structured conversation to collect personal and financial information. The solution was built on a robust architecture using AWS services, including Amazon Elastic Container Service and Amazon Aurora Serverless, ensuring low latency and natural-sounding interactions.

\## Results
After three months in production, Stretto reported significant outcomes:
- A 20% increase in lead conversion rates compared to the previous web form method.
- Tara achieved approximately 96% accuracy in data extraction during calls.
- Completed calls averaged about 8 minutes in length.
- There was a 95% reduction in per-minute voice costs compared to the previous solution.

\## Key Insights
Stretto's experience highlights the importance of validating voice quality in sensitive applications before full deployment. A phased approach, starting with a minimal viable product and expanding based on real-world data, can significantly reduce risk and improve outcomes. Additionally, leveraging advanced AI technology can transform lead generation processes in industries requiring detailed client information.

\## Customer Testimonial
“I’m incredibly grateful to the AWS GenAI Innovation Center team. Their expertise helped bring our voice agent, Tara, to life, driving a 20 percent increase in our lead conversion rate.”  
— Danya Gordin, Principal AI Architect, Stretto

## Entities

- **Companies**: Stretto, CI&T
- **Products**: Amazon Nova 2 Sonic, Amazon Bedrock, Anthropic’s Claude 3 Haiku, Twilio, Amazon Aurora Serverless v2, Amazon Simple Queue Service, Amazon Elastic Container Service
- **Technologies**: AI voice agent, speech-to-speech model, serverless computing, natural language processing
- **People**: Danya Gordin, Marcelo Wzorek
- **Organizations**: AWS Generative AI Innovation Center

## Key Concepts

AI voice agent, bankruptcy intake process, data extraction, cost reduction, voice quality validation, emotional awareness in AI, lead qualification, AWS architecture

## Definitions

- **AI voice agent**: An automated system that conducts conversations with users, often used for customer service or data collection.
- **speech-to-speech model**: A model that processes spoken input and generates spoken output in a single pass, enhancing conversational flow.
- **serverless computing**: A cloud computing model that allows users to run applications without managing server infrastructure.
- **data extraction**: The process of retrieving structured information from unstructured data sources, such as conversation transcripts.
- **emotional awareness in AI**: The ability of an AI system to recognize and adapt to the emotional states of users during interactions.

## Use Cases

- Conducting bankruptcy intake interviews
- Collecting sensitive financial data
- Sorting leads for attorneys
- Automating customer service calls
- Improving lead conversion rates
- Enhancing data quality in intake processes

## Frequently Asked Questions

**What is the main purpose of the AI voice agent Tara?**

Tara is designed to conduct bankruptcy intake interviews over the phone, collecting detailed financial information from callers in a conversational manner.

**How does Tara ensure the quality of conversations?**

Tara's voice quality is validated through testing across various emotional states, ensuring she sounds empathetic and can handle interruptions effectively.

**What technologies power Tara's functionality?**

Tara is powered by Amazon Nova 2 Sonic, a speech-to-speech model, and operates within the AWS ecosystem, utilizing services like Amazon Bedrock and Amazon Aurora.

**What benefits did Stretto observe after implementing Tara?**

Stretto reported a 20% increase in lead conversion rates and a 95% reduction in per-minute voice costs compared to their previous solution.

**How does the AI voice agent handle sensitive financial data?**

The solution incorporates AWS security and governance controls, ensuring that sensitive data is managed securely while maintaining compliance with necessary regulations.

## Links

- [Read on Welcome.AI](https://welcome.ai/content/strettos-ai-voice-agent-tara-boosts-lead-conversion-and-reduces-costs)
- [Original source](https://aws.amazon.com/blogs/machine-learning/how-stretto-built-an-ai-voice-agent-with-amazon-nova-2-sonic/)

---

Source: Welcome.AI | https://welcome.ai/content/strettos-ai-voice-agent-tara-boosts-lead-conversion-and-reduces-costs