About Valon

Valon is building the AI-native operating system for regulated services, starting with mortgage servicing. Instead of following the typical startup playbook, Valon took the “anti-wedge” path: building and operating a fully licensed national mortgage servicer from the ground up before deploying their software to other servicers. That direct experience with regulatory edge cases, exceptions, and workflow bottlenecks became the blueprint for ValonOS.

By tackling the hardest problems directly, Valon turned regulatory hurdles into lasting software advantages. Today, Valon processes over $110B in loans, delivering 60%+ margins, 90%+ customer satisfaction, and 5x efficiency compared to incumbents.

The Challenge: scaling support and maintaining trust

Valon faced a significant challenge: mortgage servicing generates high inbound call volumes that fluctuate dramatically, with peaks at the start and middle of each month and whenever new loans are onboarded. Staffing for these surges proved difficult, leading Valon to explore how AI could help them:

• Reliably handle call traffic without overstaffing
• Increase deflection rates
• Improve cost efficiency
• Deliver an exceptional, trustworthy homeowner experience

Given the sensitive and highly regulated nature of the mortgage industry, any AI solution would need to meet stringent standards for accuracy, transparency, and compliance.

How Valon chose Decagon

As Valon evaluated multiple vendors, two criteria emerged as non-negotiable:

1. Deflection rate: The percentage of calls the AI agent could successfully handle end-to-end
2. Customer experience: The ability to ensure interactions were smooth, accurate, and trustworthy among homeowners

Valon ultimately selected Decagon, confident not only in the platform's technical capabilities but also in the team's velocity of innovation and commitment to long-term partnership.

Designing a voice homeowners trust

Valon and Decagon worked closely to design AI voice agents that embodied clarity, warmth, and trustworthiness. The team invested deeply in fine-tuning the voice model to ensure it was approachable, avoided the “robotic IVR” feel, and could handle nuanced tasks such as:

• Accurate number handling - differentiating between account numbers, dollar amounts, and long sequences, delivered in the expected format at a natural pace.
• Regulatory precision – eliminating any hallucinations and ensuring financial data was correct every time.
• Rigorous QA – leveraging Decagon’s manual and automated testing frameworks to guarantee responses were always accurate.

This meticulous process gave Valon the confidence to deploy directly to homeowners.

Moving fast with self-serve tooling

Decagon’s teams partnered with Valon’s product and engineering teams to bring this implementation to life. Decagon’s ability to move quickly on improving the voice infrastructure and match Valon’s pace of innovation made them a uniquely supportive partner to Valon. In turn, Valon empowered Decagon to serve the mortgage industry by opening access to its mortgage servicing system of record and API layer, transforming its deep mortgage expertise into a platform for rapid innovation. 

Decagon’s Agent Operating Procedures (AOPs) empowered Valon’s team to self-serve - rapidly building new use cases, creating APIs, and debugging call flows. This partnership created a fast, iterative feedback loop, where both teams pushed each other to move quickly and raise the bar for AI in regulated industries.

• New use cases were spun up quickly once the foundation was in place.
• QA cycles became repeatable and automated, ensuring no regressions.
• Decagon focused on voice infrastructure improvements while Valon focused on domain expertise.

Results: Efficiency at scale without sacrificing experience

Valon’s deployment of Decagon’s AI agents has already delivered measurable impact:

• Achieved 50%+ deflection on voice for inbound calls, even when onboarding 80,000 loans in a single month.
• CSAT was maintained at ~90%, indicating that AI-driven deflection didn’t compromise customer experience.
• Average Seconds to Answer (ASA) under one minute, even during call volumes 2–3x higher than ever before.

These results demonstrate that AI can handle sensitive, regulated interactions at scale, without sacrificing accuracy or trust.

Looking ahead

Valon and Decagon share a belief that regulated industries shouldn’t have to choose between speed and trust. Valon brings its years of operating in one of the most complex financial systems in the world, along with a foundational system of record and a rich set of APIs, to power the operating system for regulated industries. Decagon leads in building AI voice agents that meet the highest bar of compliance and accuracy for sensitive, regulated customer service interactions.

Together, we are turning a historically confusing and stressful process into one that feels transparent, fast, and reliable. Homeowners get answers they can trust, and lenders gain a platform that scales without sacrificing compliance or experience.