The use of AI in clinical trials is now progressing from a stage of experimentation to one of execution. However, that doesn’t mean most AI platforms are suitable for regulated research.

In the recent episode of The Clinical Compass podcast, host Shubhangi Dua, Podcast Producer and B2B Journalist, is joined by Dr Amber Hill, Founder and CEO of Research Grid. They talk about how pharmaceutical companies and their leaders should approach the use of AI in clinical trials. Particularly, Hill points out that pharmaceutical companies should focus on AI governance, auditability and AI-native architecture rather than on flashy demonstrations or on promises of automation.

Some trends pharma companies seem to be falling for are the speed of AI adoption. Especially vendors are “attempting to adopt AI before being abreast of the technology completely and its various types of models they should be using,” states Hill.

She further says that pharmaceutical companies need to know which AI models are viable for a use case, how to train AI on the right data sets and correct information in addition to ensuring those AI models are traceable.

AI is rapidly being adopted without adequate parameters. For instance, healthcare-grade artificial intelligence (AI) and regulatory-grade artificial intelligence (AI) should be specialised AI models whose architecture is purpose-built for clinical trials and fully traceable, but some vendors with black-box AI models are using these titles without adequate checks of their architecture. Models need to be built in a way that regulatory bodies like the FDA or the MHRA can audit them to ensure a secure trail of how the model reaches its outputs.

The AI boom isn't the problem.

The problem is that many vendors are marketing AI before they've built systems that meet the standards required for regulated clinical research.

Given that pharmaceutical companies worldwide are spending billions of dollars on digital transformation, the next source of competitive advantage might not lie in using more AI, but in selecting the right AI.

Also Watch: AI in Pharma: Hype vs. What Actually Works

What does it mean to have AI-native AI and why is this relevant to clinical trials? This is the key question, Hill believes, every pharma company needs to answer before laying out the AI adoption plan.

Unlike AI wrappers, which are built on top of third-party large language AI models or other models, AI-native platforms are specifically designed, engineered and audited for the clinical trial workflow and bottlenecks they are intending to automate or resolve.

“It's not retrofitted, and the AI has learned from a model that has been built in-house for that purpose,” Hill tells Dua, alluding to AI-native models.

She explains that the main thing to consider is the AI architecture. Pharma leaders need to look into how AI models are being built, who is responsible for their creation, and whether or not the system is designed and engineered from the start to be AI-native.

Model ownership and production are another important thing to check off the list. Some questions pharma companies need to ask are: “Who has made that model? Is it a third-party consultancy company? And does that vendor actually own that model? Has it been designed and engineered in-house from the ground up?”

They need to check if the AI model has been engineered on top of open-source infrastructure like OpenAI or AWS or Anthropic or any of the other providers.

For pharmaceutical companies, that distinction has important implications.

A platform designed from the ground up provides better visibility regarding model ownership, cybersecurity, infrastructure, and audit trails. These features are essential for regulated clinical trials.

On the other hand, a great many vendors develop their products using existing 'black box' models that they do not own and about which they have only a partial understanding.

Hill says that if anything went wrong, they would have no way of knowing what had happened without getting in touch with OpenAI or Google or another owner of the original ‘black-box’AI model, who will be unlikely to backtrack the issue without Intellectual Property conflicts of interest.

“When you're thinking about traceability, you need to keep in mind if something were to go wrong, could that engineering team backtrack what's gone wrong and fix it? Or do they not own the infrastructure in its entirety to actually go back far enough to fix the problem?”

These questions are imperative to aid pharma companies in auditing a vendor that is fundamentally AI.

Overall, Hill recommends that every AI vendor should be assessed against four core areas:

• AI architecture and engineering
• Model ownership and intellectual property
• Data governance and infrastructure
• Traceability and auditability

In sectors that are subject to regulation, such as the pharmaceutical industry, these factors are becoming just as important as functionality or cost.

The question on everyone’s mind right now is if agentic AI is ready for use in clinical research. With the recent news on OpenAI’s rogue AI agents hacking into other legitimate companies' systems, such as the AI firm Hugging Face has caused the markets have been on the fence.

Agentic AI has become one of the most talked-about subjects in enterprise technology. Hill believes the industry is going too far ahead of itself. She states that agentic AI is not yet suitable for our industry since it is difficult to audit and because it has not been designed or engineered with safety and traceability in mind.

Since many agentic systems are based on generative AI and third-party foundation models, they are capable of hallucinating, of producing unpredictable outputs, and of lacking full traceability.

That presents an unacceptable risk in clinical trials.

Hill cautions that you are working with human lives and the fact that, in regulatory agencies, you may not get a second chance at that clinical trial.

“This is one of the riskiest industries there is because you're dealing with human lives, you're dealing with experimental protocols, and you're dealing with regulatory bodies where you might not get another shot at that clinical trial,” emphasises the CEO.

Rather, she thinks that AI provides the most value when it comes to solving clearly defined operational bottlenecks rather than trying to take the place of scientific decision-making.

A few examples at Research Grid are the use of specialised AI models to speed up patient recruitment, making site selection easier, or automation of paperwork to cut down on admin work.

"AI is very good at recognising patterns," Hill says. "But it has to be combined with the appropriate pain point that it is intended to address."

AI is a tool, though it does come with risks, and it is necessary to understand those risks at their core. With the speed of AI adoption in the life sciences increasing, companies which focus on AI governance, AI-native infrastructure and clinical AI that is ready for regulatory approval are likely to be in a better position.

Also Watch: AI in Pharma: Why the Future of Healthcare Starts With Patients, Not Tech

• AI-native platforms are better suited for regulated clinical trials than retrofitted AI models.
• Pharma leaders should prioritise AI governance, auditability and data ownership before investing.
• Agentic AI is not yet ready for high-risk clinical research due to traceability and safety concerns.
• Purpose-built AI can accelerate patient recruitment, documentation and clinical trial operations.
• Understanding AI architecture is becoming essential for successful pharmaceutical AI adoption.

• 00:00 Introduction to AI in Clinical Research
• 02:36 Challenges in AI Adoption for Clinical Trials
• 05:20 Auditing AI Models for Clinical Research
• 11:02 Complexity of AI in Clinical Trials
• 14:38 Industry Missteps in AI Implementation
• 17:28 Understanding Agentic AI
• 22:45 Evaluating AI Solutions for Pharma
• 27:46 Key Takeaways for Pharma Leaders

To learn more about Research Grid's AI-native clinical trial automation platform and its solutions for patient recruitment, trial operations and back-office automation, visit Research Grid.

AI vendor assessment pharma, AI vendor auditing, AI governance clinical trials, AI auditability pharma, AI-native clinical trials, AI traceability clinical trials, regulatory AI pharma, black box AI clinical trials, AI model governance, agentic AI clinical trials, clinical trial AI, pharma AI governance, Amber Hill, Research Grid

For the millions of people living with severe vision loss or complete blindness, the clinical options available today remain extremely limited. Guide dogs, white canes and adaptive technologies are designed to work around a deficit rather than address it. In a medical landscape where cochlear implants have transformed outcomes for the hearing impaired, blindness has waited for a comparable breakthrough.

That breakthrough may now be closer than most people realise. In this episode of Clinical Compass, Trisha Pillay speaks with Maarten Schelles, CTO and Co-Founder of ReVision, about the cortical brain implant his company has developed, a device that bypasses the damaged structures responsible for most blindness entirely and stimulates the visual cortex directly.

To understand what ReVision is attempting, it helps to understand why current approaches fall short. The majority of vision restoration devices on the market today are retinal implants, devices designed to stimulate remaining retinal tissue and send signals along the optic nerve to the brain. The problem is structural. These devices only work when functional retinal tissue is present and the optic nerve is intact. For patients whose blindness originates from optic nerve damage, advanced glaucoma, or extensive retinal degeneration, retinal stimulation offers nothing.

Schelles puts the scale of this gap plainly. Approximately 90 to 95 per cent of blindness cases are caused by conditions affecting the optic nerve or retina in ways that render retinal stimulation ineffective. The devices that exist serve a minority of patients with very specific pathology. For everyone else, there is no equivalent innovation which is precisely the unmet need that drove the founding of ReVision.

The cortical implant developed by ReVision takes a fundamentally different approach. Rather than attempting to restore the damaged pathway between the eye and the brain, the device bypasses that pathway altogether. Flexible electrode arrays are implanted directly onto the visual cortex, the region of the brain responsible for processing visual information and stimulate it directly to generate perception.

There are several reasons why this approach holds significant clinical promise. The visual cortex is substantially larger than the retina, which means that far more electrodes can be implanted without interference between stimulation points. Greater electrode density translates to higher resolution perception. The cortical approach also removes the constraint of requiring any residual retinal or optic nerve function, broadening the eligible patient population considerably.

The ideal candidates for this technology are individuals with no residual light perception due to conditions such as advanced glaucoma, retinal detachment, or diabetic retinopathy in its most advanced stages. Patients with intact visual cortex function but no viable pathway to transmit visual input are, in principle, candidates that no existing technology can currently serve. ReVision's device is designed specifically for them.

There are, naturally, populations for whom the approach is not appropriate. Individuals with congenital blindness, in whom the visual cortex may not have developed the expected functional organisation, and patients with tumours or structural damage to the cortex itself are not candidates. Patient selection is therefore a central component of the clinical programme.

ReVision has received FDA Breakthrough Device designation for its cortical implant, a regulatory milestone that carries practical significance beyond the recognition itself. The designation provides access to more frequent and structured engagement with the FDA during development, enabling earlier identification of potential issues and a more efficient pathway through the approval process. It also carries credibility with institutional investors and regulatory counterparts in Europe, where the company will eventually seek approval in parallel.

The clinical trial programme is structured around staged objectives, beginning with safety. Before any claims about visual restoration can be made to a regulatory standard, the evidence base must establish that the device can be implanted and tolerated without unacceptable risk. The early phases of the trial are therefore designed to generate that safety data whilst also capturing preliminary signals of efficacy, the patterns of perception that implanted patients experience when the device is activated.

Schelles is measured about what the trials can and cannot yet tell us. The full potential of the technology will only become clear as the clinical evidence accumulates. Early results from patients will shape decisions about electrode placement, stimulation parameters, and the rehabilitation protocols used to help patients interpret the signals the device generates.

The development of cortical implants for vision restoration sits within a broader trajectory in neuroprosthetics, one in which the brain is increasingly understood not as a fixed endpoint, but as a target for therapeutic intervention. The same principles that underpin cochlear implants, deep brain stimulation for Parkinson's disease, and emerging motor cortex interfaces for paralysis apply here. The brain retains a capacity for adaptation that devices can, under the right conditions, harness.

For clinicians following this space, the ReVision programme represents a great attempt to bring that principle to a patient population that has been largely overlooked by existing innovation. The regulatory pathway is defined. The clinical programme is underway. The question of real-world impact will be answered by the trial data as it emerges.

Blindness, as Schelles notes, remains a profound unmet medical need. The science now exists to address a far larger portion of it than was previously possible. What follows depends on the quality of the evidence being built. If you would like more information, visit Revision or connect with Maarten Schelles: LinkedIn | X / Twitter

• Cortical brain implants for vision restoration
• Regulatory pathway and FDA breakthrough device
• Clinical trial phases and patient selection
• Technological advantages over retinal implants
• Prospects and long-term vision

00:00 - Introduction to Vision Restoration Innovation

00:48 - Maarten Schelles' Background and Inspiration

01:06 - Origin of Revision and Brain Implant Technology

02:01 - Unmet Needs in Blindness Treatment

02:23 - Limitations of Retinal Implants

04:20 - Advantages of Cortical Over Retinal Implants

05:44 - Target Patient Populations and Ideal Candidates

07:15 - Long-term Development and Pediatric Potential

09:36 - Regulatory Milestones and FDA Breakthrough Device

12:34 - Clinical Trial Phases and Safety Objectives

19:28 - Patient Selection and Trial Design Strategy

24:23 - Insights from Early Trials and Future Strategies

27:50 - Challenges in Neuroprosthetic Development

30:04 - Closing Remarks and Future Outlook

There is a version of medical device development that treats clinical research as a late-stage obligation, a regulatory hurdle to clear before a product reaches market. Dr David Hayes has spent his career arguing against that version. As Chief Medical Officer at BIOTRONIK, he has watched what happens when clinical evidence is built into the foundation of device development, and what happens when it isn't. In a recent Clinical Compass podcast episode, Dr Hayes joined Trisha Pillay to discuss how pharma and medtech leaders should approach clinical research, collaboration, and the evolving evidence standards shaping the future of cardiac device innovation.

Clinical research in the context of medical devices is not the same as confirming a product works under controlled conditions. It is the systematic evaluation of how a device performs in real physiological and clinical environments, across diverse patient populations, under the conditions of actual care delivery.

"Clinical research evaluates devices in real environments," Hayes emphasises. The distinction matters enormously. A device that performs well in a controlled trial may behave differently when implanted in an elderly patient with comorbidities, or in a centre where workflows differ from the study protocol. Regulatory bodies, including the FDA, are increasingly demanding evidence that reflects this complexity, not just pre-market efficacy data, but ongoing demonstration of real-world performance.

For pharma and medtech leaders, this reframes the question from "what do we need to get approved?" to "what do we need to know to be confident this device is safe and effective for the patients who will actually receive it?" Those are not always the same question, and the gap between them is where post-market surprises happen.

Hayes is direct about what separates device programmes that generate meaningful clinical evidence from those that don't: the quality of collaboration between manufacturers and the clinical community.

"Feedback drives design updates and regulatory approval," he notes, and the feedback loop he is describing needs to start far earlier than most organisations do. By the time a device reaches late-stage trials, fundamental design decisions have already been made. Engaging clinicians at that point means incorporating their insights into documentation and labelling, not into the product itself. The more valuable model, in Hayes' experience, involves key opinion leaders and front-line clinicians in active dialogue during development. He recounts a recent focus group where KOL input led to substantive design changes in a product still in development. That kind of iteration is only possible when clinical intelligence is treated as a design input, not a post-hoc validation exercise. For leaders building clinical development strategies, the implication is structural; clinical affairs teams need a seat at the product development table from the outset, with mechanisms to surface and act on clinician feedback throughout the programme lifecycle.

Throughout the conversation, Hayes returns to a point that can get lost in the regulatory and commercial calculus, and that is the patient is the end-user, and their experience of the device matters. Structured patient advisory input remains underdeveloped across the industry. Gathering patient perspectives informally, or only after a product is already designed, forecloses the most impactful opportunities to improve usability, adherence, and outcomes. Hayes sees formalising patient input as one of the areas with the greatest room for improvement and, increasingly, as a dimension that regulators and payers are beginning to weigh more heavily in their assessments. For pharma and device leaders, the strategic reading of Hayes' perspective is consistent, and that is the organisations that treat clinical research as a core competency rather than a compliance function will build better products, face fewer regulatory surprises, and ultimately serve patients more effectively. For more information, visit BIOTRONIK or connect with Dr David Hayes on LinkedIn.

• Importance of clinical research in medical device development.
• Collaboration between industry and healthcare providers.
• Regulatory compliance and post-market surveillance.
• Use of real-world evidence to improve device safety and efficacy.
• Patient-centred design approaches.

00:00 - Introduction to Clinical Research and Innovation

01:00 - The Role of Clinical Trials in Cardiac Device Development

04:12 - Importance of Collaboration in Clinical Research

10:11 - Incorporating Patient Feedback into Product Design

12:10 - Navigating Regulatory Challenges in Medical Device Innovation

15:36 - Real-World Evidence and Its Impact on Device Functionality

18:54 - Strategies for Smaller Companies in Clinical Research

22:24 - The Future of Clinical Research and Patient Outcomes