Contributed Commentary by John Chinnici, CEO, Ledger Run 

January 23, 2026 | Clinical trials are foundational to all breakthrough therapies, yet they are mired in operational quicksand—particularly study startup inefficiencies. According to WCG’s 2025 Site Challenges Report, nearly one-third of respondents ranked study startup issues around contracts, budgets, and system builds as the leading problem slowing clinical trials while nearly 20% cite trial financial management (payments) as a topmost burden. 

Further, 78% of sites classified as academic medical centers and site networks, and 69% as independent sites and physician practices all agree that “contract and budget issues” are the largest contributor to long study startup timelines. Put simply, the transactional business of clinical trials is getting in the way of science, delaying new medicines from getting approved. Worse, these same issues were highlighted with the same weight two years ago in WCG’s 2023 report, signaling that no progress has been made. 

Talk about an unmet need! 

At the same time, artificial intelligence (AI) and large language model (LLM) technology has blossomed with the potential to transform clinical development. Already 80% of pharmaceutical professionals use AI to find new drugs, as numerous AI-discovered or AI-enabled drug candidates are currently in clinical development. For example, Insilico Medicine's drug for idiopathic pulmonary fibrosis (IPF), Rentosertib, is a fully AI-generated drug (both the target and the molecule were identified by AI) and has completed phase 2a trials with positive results. Molecules discovered using AI have also shown an impressive 80% to 90% success rate in phase 1 clinical trials, much higher than historic averages between 40% and 65%. 

However, in the real-world yardstick for drug discovery success—full FDA marketing approval—no AI-discovered drug has yet struck gold despite the tremendous investment. Studies show that investments cost anywhere from $25,000 to $100,000 per use case for infrastructure, development, and operational costs. Plus, given the time it typically takes to bring a drug from discovery to market, it can take many years for an investment in AI drug discovery to yield any tangible return, if ever. 

Imagine managing a portfolio of hundreds of potential AI use cases. Which ones truly align with strategic goals? Which ones offer the highest possible return? It can become unmanageable, but a data-driven approach will improve decision-making on AI investments and help ensure that resources are allocated first to high-value use cases that directly support business objectives. Right now, the industry is investing heavily in AI for high-risk, long-timeline drug discovery but largely ignoring immediate ROI in clinical operations. It's time to rebalance that equation. 

The Value Prop: Study Startup, Ongoing Operations + AI 

The technology adoption curve shows that use cases that reduce operational friction are the right place to start. Operational improvements yield quick wins, predictable returns, and accelerate adoption. Further, and in the case of AI, the latest advances (agentic AI and LLMs, in particular) are specifically tailored to automate manual workflows and read documents—ultimately, shifting work toward higher-level skills.   

Which operational areas deserve the greatest focus? Start with the greatest pain points. In clinical operations, for instance, a recent analysis indicates that about 70% of ongoing site payment operations time is spent on paper-based invoice processing. In aggregate, this translates into thousands of hours per month where an invoice is manually entered into a system to process for payment. AI can eliminate this operational waste, and then proven returns on this AI investment can support a case for future investment in other areas. 

Of course, AI cannot replace human judgment. Critical thinking is still crucial for science, strategy, and edge cases within clinical research. So, instead of focusing AI investment on high-risk use cases where lives are at stake, adopt the technology to reduce, or even eliminate, burdensome operational processes that do not directly impact patient safety. Low risk use cases in the financial administration of contracts, budgets, and payments are AI’s sweet spot. 

Clinical Trial Pain Points Where AI Shines  

1. Budget/Contract Management: One of the most onerous components of clinical research is the global negotiation of Clinical Trial Agreements (CTAs) and budgets. Each country’s unique regulatory requirements, customary practices, and negotiation processes can present significant challenges for sponsors. Local experts must collaborate closely with sponsors to develop robust, country-specific templates and fallback positions to streamline contract timelines with study sites. This process can be improved through the strategic use of AI. For example, AI can empower teams to strategically and automatically offer fallback language for faster negotiations. 

When red lines are received, modern systems can automatically scan the document and generate suggestions on alternative language within seconds, instead of manually reviewing fallback resources to determine the most effective response. AI-driven document review combined with local human experts who can apply country-specific knowledge, interpret nuances, and ensure decisions align with regulatory expectations, can significantly shorten the contract life cycle and overall negotiation time. Of course, AI does not replace the role of contract or legal professionals; rather, it can handle the labor-intensive initial review, allowing experts to focus on higher-value strategic decision-making. 

Given that clinical site contracts often take months to finalize and require substantial investment—often millions of dollars in study startup efforts—the integration of AI offers a meaningful opportunity to optimize timelines, reduce workload, and improve operational efficiency.

2. Payment Management: As mentioned, a significant volume of site payments is still triggered by highly manual, paper-based invoicing to the sponsor or CRO. The site emails a pdf document that can be 40+ pages and that describes each reimbursable activity. Someone must review the pdf and then set up proper payment to the site—often a painful back-and-forth process essentially the same as traditional paper invoicing. AI, which excels at reading large amounts of text, simplifies this human transaction.

In this example, an AI agent can receive invoice emails and identify what site, what study, and parse out all the information to load into the right places in a sponsor’s payment system, saving teams from drudging through massive documents to find what they need. Agents eliminate the middleman by reading the pdf to pre-populate a system. Of course, a human will need to be “in the loop," given the significant money transfers, so full autonomy may take time, but initial Agentic review can significantly reduce human effort. As confidence grows and the models are improved and automation will expand.  

Overall, AI can speed processes and reduce errors. Large sponsors spend thousands of hours every month processing payments. A 25% to 50% improvement can translate to millions of dollars in savings a year and multiply tenfold over the life of a multi-year, multi-site trial. 

Start With the Low-Hanging Fruit 

Sponsors have fallen into the bad habit of solving tomorrow's problems while ignoring today's fires. AI in drug discovery captures the imagination for good reason, but there's equal merit in applying AI to operational challenges that affect every trial, every day. The sites are struggling with delayed payments, the clinical research associates are pulled into contract negotiations, and the sponsors are losing competitive positioning. They cannot wait a decade for AI to mature.  

The choice is clear: continue pouring resources into AI applications that may pay off in a decade or capture measurable ROI in the next quarter. The winners will recognize that operational excellence is not just a prerequisite for innovation, but rather, the foundation that makes sustainable AI investment possible. 

Undoubtedly, AI will transform drug discovery, but can sponsors afford to wait a decade for an AI-discovered drug to mature while competitors are slashing study startup times by 30% to 40% now? Invest in parallel with operations. Build confidence. Demonstrate ROI. Then expand from there. The low-hanging fruit is not just easier to reach. It funds the ladder to everything else. Plus, investing in AI to overcome operational issues is a much safer bet with a clear, predictable return on investment.  

By using AI to automate or eliminate much of the manual processes involved in key areas like study startup (i.e., the complicated contracting, budgeting and invoicing area), everyone benefits. Sites benefit by not having to manually track and chase down payments on their invoices and can refocus on patients, helping improve patient care in clinical trials and potentially paving the road for increased trial participation in the future. And that's how we can turn AI from a buzzword into a real business advantage. 

John Chinnici is the CEO of Ledger Run. He has more than 25 years of experience running and scaling software and services companies. He has spent most of his career serving the life sciences industry, first at Accenture and IBM and then as an early leader at Veeva Systems and through much of its most rapid growth. Most recently, Chinnici was leading the Insights division of Inovalon, a healthcare software and data company. He can be reached at john.chinnici@ledgerrun.com.