Delfa is a clear example of a vertical AI agent applied to a complex, regulated industry. They are active in the application layer of the agent stack, developing agents that possess specialized knowledge of medical protocols and clinical trial logistics. Their use of voice and text as interaction surfaces highlights a trend in the ecosystem where agents are expected to perform multi-modal tasks—such as conducting an interview, evaluating responses against a set of rules, and executing a scheduling action.
For those building in the agent space, Delfa demonstrates how to bridge the gap between LLM reasoning and professional workflows. They matter because they are moving agents beyond the 'assistant' role into 'functional roles' that were previously the domain of trained staff. Their focus on protocol adherence suggests a focus on reliable, constraint-based agent behavior, which is essential for high-stakes environments like healthcare.
Clinical research is often described as a series of logistics problems disguised as medical science. The most persistent of these problems is patient enrollment. Most clinical trials fail to meet their original timelines because finding, screening, and scheduling qualified participants is a manual, high-touch process that relies on a small number of clinical research coordinators. These coordinators manage a funnel of potential candidates, many of whom are disqualified late in the process due to strict protocol requirements. Delfa is a London-based startup that addresses this inefficiency by deploying AI agents designed to own the top of this recruitment funnel.
Unlike traditional patient recruitment tools that rely on static web forms or simple logic trees, Delfa uses LLM-powered agents to conduct interactive pre-screening. These agents operate across voice, chat, and text, allowing them to meet patients in their preferred communication channel. The technical distinction is the move from data entry to reasoning. A clinical trial protocol is a complex legal and medical document with specific inclusion and exclusion criteria. Delfa’s agents are built to understand these nuances, asking follow-up questions to clarify a patient's medical history or current symptoms in the context of the trial's requirements. This reduces the burden on research sites by ensuring that only highly qualified leads reach a human coordinator.
Beyond simple screening, the company focuses on clinical trial management and scheduling. Once a patient is deemed eligible, the AI agents handle the logistics of booking appointments and maintaining engagement. This is critical for trial retention, where poor communication often leads to patient drop-off. By automating the administrative back-and-forth, the platform enables research sites to scale their operations without a linear increase in headcount. The software is designed to integrate into the existing workflow of research sites, acting as a digital extension of the clinical team rather than a standalone database.
Based in London, Delfa is an alumnus of the Entrepreneur First (EF) accelerator, a program known for backing deeply technical founders. This pedigree suggests a focus on the underlying agentic architecture rather than just a thin UI layer over existing LLMs. In the competitive landscape of clinical trial software, Delfa sits between broad recruitment platforms and specialized Site Management Systems (SMS). While others focus on trial design or data analysis, Delfa is concentrated on the execution and logistics of patient interaction. The company is currently small, with an estimated headcount of fewer than ten employees, but it targets a high-value sector where efficiency gains in recruitment have a direct impact on the cost and speed of drug development.
AI agents for automating patient pre-screening, scheduling, and engagement.
Delfa is hiring