The Experimentation Edge: How Domu’s Chief Growth Officer Uses Voice AI to Unlock New Revenue Possibilities

Picture a debt collection call where the right combination of tone, accent, and empathy turns hesitation into a firm promise to pay within seconds. For traditional call centers, such precision remains elusive at scale. Human agents vary in delivery, training cycles are slow, and testing new approaches means hiring and monitoring entire teams. 

Voice AI changes that equation. When every interaction can be controlled, measured, and iterated, the variables that once seemed intangible- warmth, urgency, familiarity- become levers for performance. 

Yet few companies treat these variables as a disciplined growth strategy. Most see voice characteristics as a product detail rather than a business driver.

At Domu, the Y Combinator S24 startup building voice AI agents for regulated financial services, Chief Growth Officer Adriana Muñoz Vergara has engineered exactly that edge.

Muñoz Vergara says: "It's the powerful combination of cutting-edge technology and rapid experimentation. At Domu, we test variables like voice tone, accent, and gender in AI calls, possibilities impossible in traditional call centers without hiring trials." 

Through relentless experimentation with variables once impossible to control in human operations, she has unlocked new revenue opportunities for Tier-1 banks and business process outsourcing firms handling loan servicing, debt collection, and customer support. 

Since joining as a founding team member, Muñoz Vergara has helped scale the company’s annual recurring revenue from $2 million to more than $10 million in under 12 months. 

Her approach treats rapid testing not as a side project but as the core of the growth strategy. It delivers 50 to 63 percent improvements in AI-handled resolution outcomes and higher customer satisfaction in deployments such as Alorica, where AI performance exceeded initial expectations.

In an industry where compliance meets innovation, Muñoz Vergara stands out by turning data-driven hypotheses into enterprise advantage.

Financial institutions and global BPOs cannot adopt AI based on technical promise alone. They require measurable ROI, regulatory confidence, operational reliability, and strict compliance. This is why Muñoz Vergara treats experimentation as an operating model itself rather than a luxury, and one early test became a defining moment.

Working with a client, Domu deployed two AI voices using the same script, timing, and collections strategy. The single meaningful difference was the voice itself. One version was immediately rejected by the customer, who asked never to be contacted again. The second voice, following the exact same process, secured a payment commitment.

Muñoz Vergara describes the approach: “At Domu, we test variables like voice tone, accent, and gender in AI calls. These possibilities are impossible in traditional call centers without hiring trials. One experiment showed different voices yielding dramatically better payment promises from the same script. This A/B testing culture accelerates product refinement, customer experience, and growth in banking operations.”

The lesson reshaped how Domu approached growth. Rather than treating voice characteristics as secondary product decisions, the company began to view them as measurable business levers that could influence recovery rates, customer satisfaction, and operational efficiency.  

Building a Growth Engine Through Experimentation

Muñoz Vergara joined Domu as a founding team member when the company served only three clients and lacked the compliance infrastructure required to compete for enterprise contracts.

Within weeks of joining, she accelerated the team's focus on conversational design. Early tests examined call flows, objection handling, escalation triggers, and script optimization for loan-servicing interactions.

Muñoz Vergara says: "We can test different voice personalities, interaction patterns, and approaches to understand what creates better outcomes for customers." 

The team had already moved into phase two, testing across Latin American markets and isolating variables such as tone, accent, gender, and pacing to assess cultural responsiveness. And the formal experimentation framework was standardized before the year ended.

Customers often reacted differently to identical messages depending on how they were delivered.

In some markets, warmer and more empathetic voices generated stronger engagement. In others, more direct and professional delivery improved outcomes. 

By testing these variables systematically, Domu transformed cultural nuances into measurable operating insights.

She explains: "One experiment showed different voices yielding dramatically better payment promises from the exact same script. This A/B testing culture accelerates product refinement, customer experience, and growth in banking operations." 

That A/B test proved voice characteristics directly affected payment promise rates. It’s a discovery that led to a formalized repeatable framework built around hypothesis creation, test design, baseline metrics, deployment, measurement, and rapid iteration.

These efforts contributed to resolution-rate improvements ranging from 50 to 63 percent. Alorica reported an NPS of approximately 50%, compared with roughly 35% for human agents.

The lesson was clear for Muñoz Vergara: experimentation transforms early traction into repeatable enterprise growth.

Why Voice Variables Matter

Behind Domu’s experimentation strategy is a deceptively simple insight: how something is said can matter as much as what is said. In debt collection and financial servicing, customers often interact under stress. Small differences in delivery can significantly affect trust, cooperation, and comprehension.

Tone influences emotional reception: a warmer, empathetic approach can encourage cooperation during sensitive discussions, while a more neutral tone may be better suited for transactional interactions.

Accents shape familiarity. Across Latin America and Brazil, Domu tested how regional speech patterns influenced engagement, helping the team understand how cultural alignment could affect customer trust and the quality of interactions.

Muñoz Vergara explains: "Being in LATAM and understanding market dynamics there further enable seamless expansion. Passion for possibility drives it all." 

In some markets and types of interactions, gender can affect how trust dynamics work. By treating gender as an independent variable, Domu was able to identify preferences that would not have been easy to detect through conventional means.

The speed of interaction affects both urgency and comprehension. Faster interaction can be useful in routine service calls, whereas slower interaction tends to be effective in hardship talks or financial talks. 

It becomes hard for traditional call centers to manage such variables, because each of them depends on human agents. Trying out a new combination of variables takes time, because one must first hire, train, and monitor the performance of human agents. 

Muñoz Vergara says: "I deliver rapid resolutions, like a 3-day banking subsidiary setup in Brazil or pricing fixes, plus systems for faster support and higher AI resolution rates."

Industry research suggests that personalized collection strategies can improve recovery performance by 15 to 30 percent. Yet many organizations still lack practical ways to personalize interactions at scale. Moreover, many AI deployments plateau at 30-40% self-service rates; Domu's continuous testing approach routinely drives that figure past 50%, as seen in the Alorica deployment.

Turning Product Insight Into Enterprise Revenue

The experimentation mindset extends beyond the product itself. Muñoz Vergara developed the commercial systems that helped transform Domu from an early-stage startup into a rapidly scaling enterprise software provider. 

She says: “I enable early-stage tech companies, especially in voice AI and fintech, to scale from traction to enterprise revenue without breaking compliance or operations.”

Muñoz Vergara personally led complex enterprise negotiations and the commercial execution of contracts with organizations including Nubank, Teleperformance, and Alorica. That work combined ROI modeling, relationship-building, and disciplined pipeline execution. 

Mark Sze, Senior Head of Product and Partner at Neon Pagamentos, a major Brazilian financial technology company that operates as a 100% digital bank,  witnessed Muñoz Vergara's approach during Domu's market entry.

He says: "She paired commercial sharpness with a real understanding of what it takes to deploy responsibly at scale. Her fluency across markets and her ability to move fast without cutting corners made our partnership possible." 

Her market-entry strategy for Colombia included financial models comparing AI-powered collections against traditional human-agent teams. By translating technical capabilities into measurable business outcomes, Domu secured conversations with Tier-1 financial institutions and moved them through complex procurement processes.

Muñoz Vergara says: "I've been heavily involved in our LATAM expansion — particularly Brazil, Colombia, and Mexico — while leading contract negotiations that safeguard IP, secure favorable terms, and prioritize long-term growth for both Domu and our clients."

The approach reflects principles she has articulated publicly: rather than leading with technology, she advocates for precise prospect targeting, deep discovery processes that address both operational and regulatory concerns, and collaborative implementation plans that reduce perceived risk.  These efforts can overcome the 38% longer sales cycles and six- to ten-stakeholder negotiations typical of AI procurement in finance. 

The philosophy of enterprise sales follows the broader trend in the AI market.

Enterprises purchase both technology and assurance. The best-selling process eliminates ambiguity by combining discovery, measurable results, compliance readiness, and implementation alignment.

Nicolas Diaz, CEO and Co-Founder of Domu, says: "Banks want effective and compliant solutions that help scale AI. Everything needs to comply with policies and leave a trace. Human-level accuracy in such a short period of time is a very important milestone, yet it marks only the start of what could be achieved." 

Case Study: Alorica Proves the Value of Experimenting

Alorica's deployment demonstrates how the experimentation framework translates into commercial results. The company is a global leader in business process outsourcing, specializing in customer experience management, technical support, and contact center operations. 

As a trusted third-party partner, the company manages customer interactions and essential back-office functions for major enterprises across the retail, telecommunications, healthcare, and financial services sectors.

Instead of treating voice AI as a fixed automation layer, Domu applied its continuous testing methodology to optimize voice patterns, conversational flow, and tone based on earlier A/B findings before the first call went live. 

In June 2026, Domu's voice AI called Rebecca processed 2,565 inbound conversations within 30 days. This AI agent resolved 48 percent of calls on its own and collected $100,000 from customers with an 87 percent success rate.

Beyond the Rebecca deployment, Domu's AI agents have managed over 16,000 conversations across Alorica's broader equipment-recovery operations.

The results showed the commercial value of optimizing what AI says and how it communicates.  

By combining automation with compliance controls and human escalation paths, Alorica improved service efficiency while preserving customer trust.

Muñoz Vergara led the commercial and integration framework for the Alorica engagement. From contract design through performance monitoring, she ensured the AI deployment met the institution's compliance and operational standards.

Jay King, President, at Alorica’s Financial arm says: “We are applying AI where it generates the greatest operational value. By handling routine, high-volume servicing tasks, the technology enables our teams to concentrate on situations that require deeper expertise and trust. The speed of adoption and performance exceeded our initial expectations.”

Operational Velocity as a Competitive Advantage

Experimentation alone does not explain Domu’s growth. Execution speed matters equally.

Muñoz Vergara’s leadership style traces back to her years playing for the Colombian national women’s soccer team. The sport taught her lessons that continue to shape how she builds organizations: urgency without panic, confidence without ego, and accountability to the team above individual recognition.

She says: "My career has always centered on finding ways to improve outcomes, not just for myself, but for the teams and businesses I'm part of." 

That mindset permeates Domu's culture. Team members frequently contribute beyond formal job descriptions, stepping into legal, compliance, operations, and commercial initiatives when needed. 

To sustain that pace, Muñoz Vergara invested heavily in automation. Using platforms such as n8n, Windmill, and Slack, she helped create automated workflows for vendor management, recruiting pipelines, candidate assessments, and client demonstrations.

These systems reduced operational friction while reinforcing a 10x operating mindset by building repeatable processes that enable team members across commercial, legal, and operations functions to contribute at a higher level rather than relying on individual effort alone.

Muñoz Vergara says: “I build systems that make everyone’s lives easier. I dedicate significant time to creating playbooks, guides, automated workflows (using tools like n8n and Windmill), and Slack integrations for vendor processing, candidate evaluation, and client demo pipelines. These replace repetitive manual tasks, allowing teams to focus on high-impact work.”

That same commitment to collective success shapes the values she instills at Domu. Muñoz Vergara says: "Extreme humility is essential for learning and growth in early-stage companies. Urgency means I ship quickly and get involved wherever needed. I believe in optimism, a 10x mindset, and constantly asking: Are we faster? Better? More ambitious? More scalable?" 

The approach has helped build long-term partnerships in highly regulated markets. And that industry perspective is regularly sought after. She has served as a judge for AI and startup competitions at institutions including UIUC, Columbia University, and DeepMind @ YC-Google. 

Muñoz Vergara's hands-on mentoring style is precisely why Omri Even‑Tov, an Associate Professor at UC Berkeley Haas, invited her as an industry mentor for the school's inaugural AI Entrepreneurship course.

Even-Tov says: "Adriana focused on reviewing the actual work product each team had built or was actively developing, which included AI product MVPs, go-to-market plans, enterprise sales frameworks, and initial pricing and revenue models for deployment. She worked as an active thought partner helping founders refine and pressure-test their output." 

Sam Tomkins, founder of Hypnos Lab and a Program Manager at Amazon, who has relied on her guidance, says: "She has advised me on defining a credible go-to-market strategy, validating the product against real customer needs rather than assumptions, and identifying where the most promising opportunities lie in the US market. Her approach is rigorous." 

For Muñoz Vergara, client success requires transparency, shared incentives, and a willingness to surface difficult truths when such conversations lead to better long-term outcomes. 

She says: “Clients should see me as someone deeply committed to their satisfaction and success beyond superficial happiness. I build relationships through honesty, even hard conversations about gaps.” 

Diaz emphasizes the breadth of her impact: "Muñoz Vergara operates across commercial strategy, regulatory compliance, and technical execution. These are functions that most organizations staff separately. It’s the combination of skills that enables us to experiment, iterate, and scale at the pace we do."

As Domu expanded internationally, Muñoz Vergara helped oversee the establishment of the company’s Brazilian subsidiary and contributed to a licensing effort that expanded Domu’s U.S. footprint from fewer than three licensed states to more than 35. At the same time, the company strengthened compliance capabilities across frameworks, including SOC 2, PCI DSS, ISO 27001, GDPR, and LGPD.

The strategy also produced direct economic benefits for customers. In Brazil, Muñoz Vergara led telecommunications pricing negotiations during a period when billing changes and cost pressures threatened operational efficiency. By working directly with providers and restructuring pricing terms, she helped reduce losses by tens of thousands of dollars per day while restoring healthier margins for the business. 

She highlights the value of challenges in her work: “It sharpened my skills in rapid crisis response, regulatory navigation, contract strategy, and cross-functional leadership.” 

Voice AI as a Continuous Optimization Engine

The next chapter for Domu centers on extending experimentation beyond inbound and outbound servicing.

This roadmap expands the role of voice AI from reactive servicing to proactive customer engagement. Outbound reminders, early intervention strategies, and personalized financial guidance can become continuously optimized revenue channels.

The company is developing outbound voice AI capabilities designed to support payment reminders, delinquency prevention, and personalized financial guidance. Future deployments will combine AI voice interactions with SMS and email outreach. This will create coordinated multi-channel engagement strategies optimized through continuous testing. 

At the same time, she is aware that AI-driven quality assurance systems will automatically evaluate conversations for compliance, empathy, effectiveness, and risk indicators.

Muñoz Vergara says: "I’m excited to ensure expansion efforts, market research, and outreach, so they align with our ideal customer profile, maximizing ROI on new initiatives.”

Her broader vision is ambitious but consistent with Domu’s philosophy: “Voice AI should not function as a static technology deployment. It should operate as a learning system that improves with every interaction,” Muñoz Vergara adds.

Industry analysts increasingly view the speed of experimentation as a defining advantage in enterprise AI adoption. The companies that can test, measure, and refine customer interactions faster will have a greater opportunity to build durable advantages in regulated markets.

While many competitors focus primarily on language understanding or infrastructure performance, Domu has added another layer of optimization by treating voice characteristics themselves as measurable business variables.

Muñoz Vergara explains: "Results like these, faster feedback, higher satisfaction, and operational gains, exemplify how I bridge commercial strategy, product, and client success."

However, the mission extends beyond growth metrics. Muñoz Vergara says: “As someone who turned ‘impossible’ into reality through technology experimentation, openness to new possibilities, and relentless drive to open doors for growth. A discovery-oriented builder who learns from every test.” 

With the continued development of voice AI, leaders like her prove that experimentation is a product development tool, a business strategy, an operating model, and a competitive moat.  

Her example shows how to create high-performing organizations: combine fast experimentation, strong operations, clear collaboration, and the ability to implement solutions. 

“The companies that learn fastest will win in this age of artificial intelligence," says Muñoz Vergara.

“I push for better outcomes for teammates, clients, and projects, expanding possibilities personally and professionally through experimentation and full commitment. This helps businesses achieve their best.”

Muñoz Vergara's approach, treating experimentation as the operating system instead of a project, offers a blueprint for that speed. 

About the Author