Vertical Relevance partnered with Numerix, a global financial services technology company, to architect, engineer, and deploy Da Vinci, a governed AI development platform on AWS with Anthropic Claude as its reasoning layer.
The platform provides Numerix with a reusable foundation for building and deploying AI agents and agentic applications while establishing consistent standards for security, permissions, observability, tool integration, and usage and cost management. A production Claude powered agent built on the platform is helping Numerix customers construct API payloads for quote and pricing requests, with customers reporting that the process now takes approximately 10% of the time it previously required an estimated 90% reduction.
The Business Challenge
As interest in AI grew across Numerix, teams faced a common challenge moving AI concepts into production required repeatedly solving the same foundational questions around infrastructure, security, permissions, observability, model integration, and tool access. Numerix needed a common framework that could accelerate AI development while maintaining the governance, security, and auditability required in financial services.
Speaking with theCUBE during the AWS Financial Services Symposium in May 2026, Satyam Kancharla highlighted the opportunity for AI across capital markets alongside the importance of governance, security, and auditability as these capabilities move into production.
The Solution: AI Development Platform with Anthropic Claude
Vertical Relevance partnered with Numerix and architected, engineered, and deployed an AI development platform that standardizes how Numerix builds AI infrastructure, agents, and agentic applications.
At the center of the platform is Anthropic Claude, providing the reasoning capabilities that power Numerix’s AI agents. Da Vinci, an AI development platform creates a repeatable framework around Claude so teams can focus on business use cases rather than rebuilding the underlying foundation for each application.
Key capabilities include:
- Reusable infrastructure: AWS CDK constructs provide standardized building blocks for deploying Claude-powered applications.
- Configuration-driven development: Developers define application requirements through a manifest, allowing Da Vinci to generate standardized infrastructure and application patterns.
- A shared Claude-based AI SDK: Built on Strands, the SDK provides common capabilities for Claude, including inference handling, usage and cost tracking, permissions, and tool integration.
- Standardized tool integration: Model Context Protocol (MCP) patterns provide a consistent way to connect Claude-powered agents with enterprise data, documentation, and tools.
- Flexible agent architectures: The platform supports single-agent, multi-agent, and broader agentic applications
.
As Brian Jakovich explained on theCUBE, the goal was to create a foundational platform that allows Numerix to take a business use case from concept to implementation and then apply the same foundation to additional use cases.
Claude Powered Agent for Quote and Pricing Requests
A production application built on Da Vinci is a Claude-powered knowledge and support agent for customers of Numerix’s SaaS products.
Numerix customers use these products to submit structured requests for quotes and pricing on financial instruments. Because the APIs rely on complex request schemas, customers previously had to navigate extensive technical documentation to determine the appropriate fields and payload structure.
The Claude powered agent simplifies this process. Using an MCP-based documentation capability, it retrieves relevant product documentation and uses Claude’s reasoning capabilities to interpret the customer’s request, identify the appropriate schema and required fields, and help construct a correctly structured API payload.
Outcomes
- An estimated 90% less time building payloads. Customers report that building a valid API payload with the agent takes about 10% of the time it used to.
- One production foundation. New agent use cases start from established infrastructure and software patterns instead of repeating model integration, security, tool access, observability, permissions and cost controls.
- Shorter path to production. Application teams move from an AI concept to a governed implementation faster, with more sophisticated agentic patterns within reach.
- Claude as shared infrastructure. Claude is a shared reasoning layer that production applications can use through repeatable, governed building blocks, not a one-off feature or isolated chatbot.
Looking ahead: the RFQ Agent
During the same theCUBE interview, Kancharla also discussed an emerging RFQ Agent that Numerix was developing with customers.
The concept is designed to combine Numerix’s established deterministic models with AI reasoning to help traders and sales professionals respond to complex client requests involving requirements such as liability, KYC, and risk constraints.
The agent is intended to interpret those requirements and constraints and use them alongside Numerix’s existing models to help tailor a quote to the specific client need. Kancharla described the opportunity as enabling financial institutions to provide a level of service that historically was available primarily to their largest and highest-priority customers.
Why Vertical Relevance
Vertical Relevance combines deep financial services expertise with AWS and AI engineering capabilities to help financial institutions and financial technology companies move AI from experimentation into production.
For Numerix, that meant going beyond a single AI application to establish a governed, reusable foundation for Anthropic Claude-powered innovation—one that connects business requirements to production technology and creates a path for additional agentic applications over time.
As Jakovich described on theCUBE, the opportunity is increasingly about applying technology where it can create measurable business value: accelerating work that previously required significant manual effort and bringing AI capabilities closer to the business processes where they matter most.