By Deborah Borfitz

June 17, 2025 | An artificial intelligence (AI) agent known as Grace is bringing clinical trials some much-needed speed and cost reduction and delivering more personalized, on-demand digital experiences for participants. Developed by fast-growing Grove AI, Grace has over the past year powered more than 50 phase 2 and 3 trials involving over 500,000 interactions with would-be and enrolled trial participants, shares company cofounder Sohit Gatiganti. 

The communications include phone calls, text messages, and emails to accomplish tasks such as screening participants and then making and reminding individuals about their on-site screening appointments. But Grace is not just another AI-powered call agent for automating patient recruitment and screening, he says. Grove’s AI Participant Relationship Management works “behind the scenes” to surface critical trial insights in real-time, enabling research teams to make data-driven actions on screening and enrollment. 

IMA Clinical Research, a site network comprising 17 locations nationally, recently piloted the use of Grace at seven of its sites for a phase 3 cardiometabolic trial for a large pharmaceutical sponsor, reports Mark Weinberger, Ph.D., president and CEO of both IMA Clinical Research and its parent, The IMA Group (focused on medical and behavioral health evaluation services). Grove AI did “a phenomenal job” with enrollments in the first 30 days, and the company is now expanding the collaboration to other indications and therapeutic areas such as psychiatric and central nervous system conditions. 

In addition to being compliant with federal laws (i.e., Health Insurance Portability and Accountability Act) and regulations (i.e., CFR 21 Part 11), Grove was the first to receive generic IRB approval from one of the leading institutional review boards—meaning, Grace can be integrated into clinical trial workflows  more quickly than other AI platforms without sites needing to include additional technical descriptions of the AI in study-specific material submissions. 

Gatiganti notes that Grove’s success is grounded in its ability to engage participants exactly when they are most likely to respond, 24-hours-a-day, seven days a week. Initial prescreening calls happen instantly when individuals express an interest in a clinical trial after learning about it on social media, he says. Outbound prescreening calls are also made to people who have previously registered their interest. 

The messaging is tailored to each study, according to the approved script provided by sites using their branding. “We do well across diverse conditions and patients [because] we adapt to people’s needs,” says Gatiganti. With the Alzheimer’s disease population, for example, Grace will speak slower and repeat herself more. “Some of our patients are on the phone with us for an hour just talking,” he adds. 

If the participant asks the agent to call back at a more convenient time, Grace will also reach back out at the precise day and time as requested by that participant, Gatiganti continues. To respect the time constraints of participants and study sponsors, as well as save sites hours of manual data entry, Grace is being built as an end-to-end tool that can be integrated with any clinical trial management system. 

Grove AI is focused on patient recruitment and engagement because the biggest bottleneck in clinical research today is not lead generation but lead activation, Gatiganti says. Patients interested in trial participation need support “in real time, in any language” through various stages of a study. 

Piloting Grace 

IMA Clinical Research is conducting dozens of phase 2 through phase 4 clinical trials at any one time in a variety of therapeutic areas, says Weinberger, and regularly meets vendors who can potentially help it streamline the trial process, enhance efficiency, and improve the quality of research. “The only way clinical research businesses make money and advance science is bringing subjects into the office and performing trials,” which is why IMA proactively reaches out to patient recruitment companies. 

The partnership with Grove AI began with an introduction to the company by a professional colleague, he says. But it is likely that IMA Clinical Research would have eventually crossed paths with the startup given its visibility at industry conferences and on social media. 

The initial pilot study with Grove AI focused on the impact of linoleic acid on cardiometabolic health to test out the company’s capabilities and economic model, says Weinberger. Individuals in an internal database and others identified via ads on social media were “quickly and efficiently” called by Grace, following a script of questions created by IMA.   

If their responses affirmed their prequalification for the trial, Grace tried to schedule them for an in-person screening at one of the seven study sites, he explains. Grace sounds so much like a human that although she introduces herself as a digital assistant, it’s likely not everyone realized they were talking to AI machinery. If participants requested a human, or needed medical assistance, they were immediately transferred to an actual human—an escalation rate that Gatiganti pegs at well under five percent. 

Good ROI

The results of the pilot study were encouraging from several perspectives, Weinberger says. For starters, it generally takes site staff at least a couple days to reach out to anyone who expresses interest in a trial via a form they submit through the IMA website, versus being called by Grace within seconds of that click. “It’s playing off of a captive audience.” 

Grace succeeded in accelerating prescreening for the trial by triaging over 1,500 participants in 30 days. The most obvious return on investment is in the accelerated timelines, reduced costs and averted staff burnout as Grace, not IMA employees, are making all those calls, says Weinberg. 

Among the on-site appointments made by Grace, there was about a 60% show rate, which is “in line with... if not slightly better” than that of its human counterparts, he adds. The enrollment rate of people engaged by Grace was a respectable 32%—a figure IMA Clinical Research hopes to improve substantially by perfecting its screening questionnaires. “That’s on us.” 

That’s to say nothing about the savings associated with not having to hire local recruiters (an estimated three to four for the pilot study) across the site network. Grace hit the trifecta in terms of desirable speed, accuracy, and cost outcomes, says Weinberger. 

Building the Database 

In addition to accelerating existing clinical trials, Grace is powering efforts to set the foundation for upcoming research opportunities. IMA Clinical Research just launched another trial with Grove AI, this one a gastrointestinal preventive screening study. Grove AI is helping to build up the pool of willing study participants for many of IMA’s sites, Weinberger says. This will give them more of a “fully vetted patient database” to tap into when enrolling people in future clinical trials.  

“Given the changing technology landscape, I think clinical research sites need to be open-minded about new ways of doing things,” says Weinberger in talking about the partnership with Grove AI. Use of the clinical research agent is certainly a better approach to diversifying and advancing science than leaving people on their own to come in for a screening once a study has piqued their interest. 

“In today’s digital age where AI adoption is increasing, people are getting super-comfortable with the use of AI—not just enterprises but consumers in general,” adds Gatiganti. “We just need a faster, more intelligent way to operate clinical trials, both for participants and patients to get them the care and attention they need but also for sites to accelerate the manual workflows and deliver better trials overall.”