EClinCloud launches its embedded AI agent and next-generation IRC platform
This English editorial adaptation summarizes coverage first published by Artery Network on May 20, 2026.
At its 2026 user conference in Shanghai, EClinCloud introduced an embedded clinical-research AI agent and presented the latest generation of its independent imaging review platform.
Founder and CEO Michael Qin, PhD, described the industry's progression from large language models to multimodal systems and then to agents that can participate in controlled workflows. The company calls this direction AI for Clinical: applying AI to concrete study roles while preserving professional review, traceability, and regulatory accountability.
A four-year AI development path
EClinCloud's applied AI program began in 2022, before the current wave of generative AI adoption. Early work included a knowledge-graph clinical chatbot developed with a multinational customer to support conversational assessment of patient eligibility criteria.
The company later established a dual path of proprietary models and large language models, moved from algorithm development to production engineering, and built a clinical-trial RAG knowledge base with structured prompt engineering. In 2026, those capabilities became embedded across ECC platform workflows.
EClinCloud now reports more than 20,000 users, 250 organizations, and 1,100 clinical trials across more than ten countries and regions. Product workflows include study build, source-document recognition, data review, operational monitoring, document management, and multilingual site support.
Lin Xiaofu: assistance shaped around the role
The new agent is designed as a set of role-specific assistants rather than one general-purpose interface.
For study build, it can parse a protocol and draft visit schedules, forms, and validation logic for expert review. For site users, medical OCR and multimodal models help turn source documents into structured data and flag privacy-sensitive content. For project teams, predictive models can surface enrollment, compliance, follow-up, and other operational risk signals. A multilingual support layer provides guidance across global sites and time zones.
Every production use remains governed by permissions and human review. The objective is to remove repetitive configuration and search, not to transfer accountable clinical judgment to a model.
A broader independent imaging review capability
EClinCloud also demonstrated its IRC platform, which connects image upload, anonymization, quality control, assignment, expert reading, adjudication, progress monitoring, and reporting.
The platform supports DICOM 3.0 and additional imaging and document formats and is expanding from ophthalmology into cardiovascular, autoimmune, endocrine, oncology, neurological, and other applications. The company is also exploring AI-assisted lesion segmentation and the preparation of enrollment, imaging, and consistency reports.
Building an open clinical AI ecosystem
The conference included leaders from pharma, biotech, CRO, AI drug discovery, and health technology. Their discussion focused on a shared challenge: clinical AI must improve speed without weakening the quality and governance of the data it produces.
EClinCloud's stated direction is to work with technology and clinical partners on an open, sustainable ecosystem that connects platforms, real study scenarios, and professional services. The goal is practical adoption — AI that operates inside a reliable clinical evidence chain.
