Florence, a clinical trial technology provider known for pioneering the electronic investigator site file (eISF), is positioning AI as a way to connect fragmented clinical research workflows rather than simply collect more data. At its Research Revolution event in Atlanta, CEO Ryan Jones outlined a strategy that combines AI copilots and agentic capabilities to automate tasks between research sites and sponsors while keeping people responsible for decisions. The approach targets administrative work across study feasibility, document quality, study startup and monitoring, as clinical research organisations face growing demands on site capacity.
Florence Targets Clinical Trial Bottlenecks With AI Agents
Clinical research relies on a network of sponsors, investigators, hospitals, research sites and technology vendors. Yet the systems supporting a trial often operate separately, leaving research teams to transfer documents, re-enter information and reconcile processes across electronic consent, patient-reported outcomes, electronic health records, imaging and study-file platforms.
Florence wants to change that model. At its Research Revolution event in Atlanta, the clinical trial technology company outlined a strategy built around connecting workflows and automating routine work, rather than adding more systems for research teams to manage. Its vision, described as “connect, not just collect,” puts AI copilots and agentic capabilities to work across tasks performed by study sites and sponsors.
The company says the goal is to let software handle more of the administrative work while researchers retain control over decisions. That distinction is particularly important in clinical research, where document accuracy, regulatory compliance and human oversight can directly affect study operations.
A fragmented technology landscape
Clinical trial technology has expanded to support more digital processes, but individual systems do not always communicate smoothly. Study teams may need to respond to sponsor questionnaires, prepare regulatory documents, verify incoming files and share updates through email or separate portals.
Florence cites its 2026 State of Clinical Trial Technology Report, which found that 64% of surveyed sites said sponsor technology duplicated work they already performed. The report also found that 95% still exchanged study documents through email. These are company-reported survey findings, rather than independently verified measures of the entire clinical research sector, but they illustrate the operational friction Florence is targeting.
For sites managing multiple studies, repeated data entry and document handling can compete with the time needed to coordinate investigators, support participants and keep trials on schedule. Sponsors also depend on sites to provide consistent information across feasibility assessments, study startup and ongoing monitoring.
Florence argues that adding more software without connecting existing workflows can compound the problem. Its strategy instead focuses on automating the movement and review of information across tasks that have traditionally required manual intervention.
Copilots and agents for routine research tasks
Florence says it is already applying this approach through two products. Site Feasibility, which the company says went live in September, draws on a study site’s library of previous answers to help complete sponsor questionnaires. AI recommends responses, while a person makes the final decision.
DocQC checks incoming documents for quality issues. Identifying problems as documents arrive could help teams resolve missing or incorrect information earlier, instead of discovering issues later in a study process.
These examples illustrate different roles for AI in clinical operations. A copilot can recommend an answer or help a user complete a task; an agentic workflow can coordinate more of the process across systems and steps. Both approaches still require clear boundaries around what software can change, submit or approve.
Florence says it will share further details about AI workflows for feasibility, study startup, document quality and monitoring on October 27. The announcement does not provide quantified results for the two products, such as reductions in processing time, error rates or staff workload. Those measures will be important for assessing whether the technology improves research operations in practice.
Capacity pressures put focus on efficiency
Florence also frames workflow automation as a response to a growing gap between the development of potential therapies and the capacity of research sites to run trials. The company estimates that pipeline drug candidates are increasing by nearly 8% annually, while site activations are growing by around 4% annually. Based on those figures, Florence estimates a 33% gap between therapies ready for testing and sites available to conduct studies.
These figures are Florence’s estimates, and the announcement does not provide enough methodology to independently assess the calculated gap. Still, the underlying challenge is relevant: expanding drug development does not automatically increase the number of sites, investigators and research staff available to conduct studies.
Reducing administrative work could help sites use their existing resources more effectively. However, software automation alone cannot resolve every cause of limited trial capacity, which can also involve recruitment, staffing, infrastructure, study complexity and the suitability of sites for particular protocols.
From system of record to system of automation
Florence introduced its eISF in 2015 as a digital system for maintaining study documentation. The company now describes a progression from systems that store information to systems that connect tasks and, ultimately, automate parts of the work itself.
Its scale provides a foundation for that transition. Florence says its platform supports 88,000 study sites across 90 countries and has been used to create more than 44 million documents. The company also reports record study volumes exceeding 50,000 protocols and support for approximately 140 regulatory approvals over the past decade, including 84 medical device approvals and 39 drug and biologic approvals.
These figures describe the company’s reported platform footprint and historical activity; they do not establish how many sites currently use its AI products or how much work those products automate.
Human oversight remains essential
Clinical research presents particular challenges for AI deployment. Incorrect document assessments, unsuitable feasibility answers or poorly managed workflow changes can introduce delays and compliance risks. Automation therefore needs to be transparent enough for research teams to check outputs, correct mistakes and understand when escalation is required.
Florence’s stated model keeps people responsible for final decisions in the workflows it has described. As the company expands agentic capabilities, research organisations will need clarity on data access, audit trails, validation, permissions and how automated actions are reviewed.
Florence’s strategy reflects a wider enterprise AI shift toward coordinating work across existing systems rather than simply adding a chatbot to each application. In clinical trials, the opportunity is to reduce operational friction without transferring scientific, clinical or regulatory accountability away from people. The test will be whether connected workflows produce measurable efficiency gains while maintaining the reliability and oversight that clinical research demands.
Market Landscape
Clinical research technology is moving toward connected workflows and AI-assisted operations as sponsors and research sites look to reduce administrative burdens. Florence’s strategy focuses on workflow orchestration rather than replacing individual clinical trial systems.
- Fragmented trial systems: Research teams often work across eConsent, EHR, ePRO, imaging and electronic study-file platforms, creating manual coordination work.
- AI-assisted site operations: Feasibility questionnaires and incoming-document quality checks are practical early use cases for AI recommendations and automation.
- Agentic workflow development: Vendors are exploring ways for agents to coordinate multi-step tasks across systems while retaining human approvals.
- Regulatory accountability: Clinical trial automation requires validation, traceability, appropriate access controls and clear responsibility for decisions.
- Capacity constraints: Reducing repetitive administration may help research sites use existing resources more efficiently, although it cannot independently resolve every trial-capacity challenge.
Top Insights
- Florence is shifting its clinical trial technology strategy from document collection toward connected workflows and AI-assisted execution across sites and sponsors.
- Site Feasibility uses a library of previous site responses to recommend answers to sponsor questionnaires, with people retaining final control.
- DocQC checks incoming documents for quality issues, aiming to help research teams identify problems earlier in study workflows.
- Florence reports supporting 88,000 study sites in 90 countries and creating more than 44 million documents on its platform.
- The impact of its agentic strategy will depend on measurable efficiency gains, document accuracy, integration quality and regulatory oversight.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
