ACORD adds MCP architecture to its insurance AI suite

ACORD Solutions Group moved its insurance AI pitch from aspiration to architecture on May 28, introducing Model Context Protocol-enabled capabilities across ADEPT and ACORD Transcriber. The company’s case is simple but consequential: insurers do not need to rip out core platforms to test agentic AI if a trusted execution layer can sit between agents and the systems that already handle policy, claims, document, and workflow work.

That is the real question behind the launch. In software terms, MCP only matters if it reduces the integration bottleneck that has slowed insurance automation for years. ACORD says it does by giving AI agents a standards-based way to request, validate, reconcile, and route data across its digital insurance suite, rather than behaving like a standalone chatbot layer. Chris Newman, president and CEO of ACORD Solutions Group, called the release a “defining moment” for the industry.

The examples ACORD chose point directly at operational orchestration. A broker-side agent can request and validate a digital invoice. An insurer-side agent can help reconcile that invoice and prepare it for processing. A claims-handling agent can pull loss data from incoming documents. A placing agent can coordinate quote, firm order, and bind-ready data while human underwriters keep approval authority. That is a narrower and more useful promise than generic AI assistance: it suggests AI acting inside existing workflows, not outside them.

The strength of the approach depends on the quality of the plumbing underneath it, and ACORD has spent years building that plumbing. ADEPT, ACORD’s data exchange and translation platform, is built for real-time data exchange, structure, validation, and reconciliation. ACORD Transcriber can extract data from more than 4,700 versions of more than 800 ACORD forms, with human-in-the-loop review for potential inaccuracies. ACORD has also said ADEPT supports more than 50 ACORD Transcriber AI models, tying document extraction directly to downstream data movement.

That broader strategy was already visible earlier in the year. In January, ACORD argued that agentic AI could support data analysis, routine decisions, and data enrichment in underwriting and claims analysis. In August 2025, it made the same underlying point more bluntly: insurance data standards are vital if AI is going to work effectively. The May 28 launch pushes that thesis one step further by treating standards as a prerequisite for autonomous workflows, not just a nice-to-have for interoperability.

There is also a market signal behind the messaging. ACORD said in May that Aon completed an integration with ADEPT to accelerate placing transactions, giving the company a live example of enterprise adoption. At the same time, ACORD launched Core Data Record v3.3 to support treaty reinsurance, underscoring how closely it is tying AI ambitions to data standardization across the industry.

The release does not eliminate the hard parts insurers still face. Carriers will still need clean source data, clear approval rules, and disciplined orchestration before agents can work reliably at scale. But ACORD’s bet is clear: the next phase of insurance AI will be judged less by model novelty than by whether it can move structured work through the enterprise with auditability, control, and minimal manual intervention.