dv01, a capital markets fintech company focused on technological innovation and loan-level transparency in structured finance, has unveiled its agentic infrastructure. The launch covers AI agents for the company's DealStudio and Credit Facility Management offerings, along with Model Context Protocol (MCP) connectivity for collateral analysis and cashflow modeling. Together, these capabilities enable firms to apply AI across asset-backed securities (ABS) and residential mortgage-backed securities (RMBS) workflows — markets in which analysis depends on the behavior of large pools of underlying loans.

“The industry is approaching a fundamental shift in how work gets done, and AI cannot be treated as a feature. Agents need transparent loan-level data, reliable analytics, and fuller context of a transaction to connect insights that may be fragmented across teams and systems,” said Jonathan Warrick, Head of dv01. “We’ve spent more than a decade building that foundation and are adapting our infrastructure to support agent-driven workflows, so firms can move beyond automating tasks and expand what is possible across structured finance without lowering the for rigor, control, and accountability.”

Operationalizing AI with Built-In Agents and MCP

dv01's agentic capabilities reach the market through two channels.

AI agents built into the dv01 platform: Users provide source documents and instructions, then review, refine, and approve the resulting work — a structure that keeps human judgment in the loop. The agents can, for example, turn an offering memorandum into a draft structure, or use a credit agreement and loan data to configure a facility and draft borrowing base and monthly servicer reports.

MCP for client-built agents: Firms can analyze collateral and run cashflow projections using dv01 data and analytical capabilities within their own AI environments. ChatGPT, Claude, or internal AI applications can connect through the dv01 MCP server without building a separate custom integration. MCP itself is an open standard, introduced by Anthropic, that standardizes how AI applications connect to external data sources and tools.

Semantic Layer Grounds AI in Structured Finance Context

The dv01 Semantic Layer assigns consistent business meaning to standardized loan-level data by mapping source fields to defined concepts, establishing how metrics are calculated, and incorporating documented deal- and servicer-specific exceptions. That grounding addresses a practical reality of structured finance, where fields and metrics can be defined or reported differently across deals and servicers. Through the dv01 MCP server, large language models (LLMs) call dv01 capabilities built on this foundation instead of interpreting raw data or recreating methodologies within the model. According to the company, this gives agents the structured finance context needed to return analysis grounded in dv01-defined data and methodologies.

Building Agentic Infrastructure for Structured Finance

AI has become a strategic priority across capital markets, but deployment remains uneven across firms and desks. In dv01's view, as agents take on more of the preparation and execution of work, market participants will increasingly focus on defining objectives, directing workflows, reviewing evidence and exceptions, and applying judgment — a shift that makes the infrastructure beneath the work all the more important. The company positions itself as the partner providing the infrastructure to integrate agents into established workflows, grounded in the data, calculations, domain context, and controls the market requires.

Built-in AI agents are available within their respective dv01 product lines, and MCP access is live for Claude and ChatGPT, with plans to expand to other AI platforms. Additional information about the company is available on dv01's official website, and the original announcement was published by Global FinTech Series.