Digital Science has rolled out two new Model Context Protocol (MCP) server products for its Dimensions research intelligence platform, enabling enterprise AI agents to connect directly to live, licensed research data.
What the new connections do
The company says the integrations allow AI-powered assistants to query a globally linked database of research outputs without relying solely on general-purpose training corpora. Dimensions already aggregates more than 430 million interconnected records covering publications, grants, patents, clinical trials, datasets and policy documents. The new MCP endpoints are presented as a route for organisations to automate discovery, competitive intelligence and funding analysis inside AI workflows while retaining access to up-to-date, verifiable sources.
| Product | Primary purpose |
|---|---|
| Dimensions Semantic Search MCP | Concept-based retrieval across research domains, especially in life sciences |
| Dimensions Analytics MCP | Linked-data access for mapping research landscapes and performing intelligence analytics |
Why it matters
Enterprise AI systems frequently fall back on broad, static training datasets that can be outdated, incomplete or hard to verify. By exposing Dimensions through the emerging MCP standard, Digital Science aims to let AI assistants fetch structured, live records for evidence-based answers and data-driven workflows. The company highlights that MCP is already supported by several prominent AI platforms, which should make integration into existing enterprise stacks straightforward.
- 430m+ interconnected research records are available via Dimensions.
- Semantic retrieval is tuned to recognise concepts rather than simple keywords across more than 40 life-science subject areas.
- Current Dimensions API customers can use the MCP servers without an extra licence, the company says.
Implications for research and policy
For universities, research funders and commercial R&D teams, the capability could shorten the time between posing a question to an AI and obtaining a traceable, data-backed result. Use cases include mapping active research areas, profiling institutions and investigators, analysing funding trajectories and linking outputs to people and organisations. That connectivity is particularly valuable where decisions require verifiable provenance—grant allocation, regulatory scrutiny and technology scouting among them.
There are also governance questions. Linking enterprise AI assistants to licensed databases changes who controls access to curated research knowledge and how that access is logged and audited. Organisations will need to consider licence terms, data security, and whether AI responses retain sufficient citation metadata to support reproducibility and accountability.
How the MCP approach differs
Rather than embedding static snapshots of papers and metadata inside a model, the MCP approach operates more like a live query gateway. The Semantic Search product is described as optimised for idea-level retrieval—identifying relationships between drugs, diseases and compounds—while the Analytics product surfaces linked metadata for analytics workflows.
Digital Science says existing customers of the Dimensions API can begin using these MCP servers immediately. The move reflects a broader industry push to make AI assistants less brittle and more verifiable by tethering them to authoritative, structured data sources.
As organisations increasingly stitch AI into decision-making pipelines, tools that preserve provenance and permit real-time queries of licensed research databases will be judged not just on speed, but on the trustworthiness of the answers they produce.
