AI’s Impact on Small Biotechs: Why Better Decisions, Not Just Faster Processes, Will Define Success

Authored by

AI in Clinical Development: Why Better Decisions, Not Faster Processes, Will Define Small-Biotech Success.

Artificial Intelligence (AI) is reshaping drug development, but its impact may look very different for emerging biotechs than it does for large pharmaceutical companies. While major pharma organizations invest heavily in proprietary AI tools and infrastructure, smaller companies are exploring how AI can help lean teams move faster, make smarter decisions, and compete more effectively with limited resources. 

To explore what this means in practice, we spoke with Aiden Flynn, Senior Vice President of Strategic Statistical Services, and Irving Dark member of MMS Board of Directors, about how AI is changing clinical development, biometrics, trial design, and sponsor-CRO relationships.  

Is AI Creating a Fundamentally Different Opportunity for Small Biotechs Than for Large Pharma?

Aiden Flynn: Absolutely. Large pharmaceutical companies are investing enormous resources into building their own AI capabilities and infrastructure. Smaller biotechs typically don’t have that luxury. Instead, they’re looking to leverage available tools and external expertise to help them move faster and operate more efficiently. 

What’s interesting is that we’re already seeing sponsors arrive with work that has been generated using AI. In one example, a company presented us with a statistical analysis plan that had been developed using AI-assisted methods. A few years ago, that would have been unusual. Today, it’s a glimpse of where things are heading. 

Irving Dark: Large pharma and biotechs are approaching AI from very different positions. Large organizations are primarily focused on cost reduction, productivity gains, and accelerating timelines. Small and mid-sized biotechs have a different opportunity. AI has the potential to help them do more internally, gain greater visibility into their programs, surface insights in real-time, and become more informed consumers of outsourced services. 

Over time, we may see a fundamental shift in how biotech companies interact with service providers because they’ll have access to information and analytical capabilities that weren’t previously available to them due to budget constraints. 

How Could AI Change the Relationship Between Sponsors and CROs?

Irving Dark: Historically, there has been a fairly significant gap between sponsors and vendors. What we’re seeing now is that gap starting to close. AI-enabled tools are creating more visibility and more opportunity for collaboration. Rather than simply transferring responsibilities to a CRO and waiting for outputs, sponsors are becoming much more involved throughout the process. 

Outsourcing will continue to evolve. What changes is how sponsors and vendors work together. The future is likely to involve more shared responsibility, more transparency, and a closer working relationship built around enhancing decision-making rather than just deliverables. 

Aiden Flynn: That shift also changes where value is created. For years, many CRO business models have been dominated by implementation services and the resources needed to perform them. AI will increasingly automate portions of that work. As a result, knowledge-based services, strategic guidance, and expert oversight become even more important. 

The balance may not flip completely, but it certainly won’t remain where it is today. 

Where do you See the Greatest Opportunity for AI in Clinical Development?

Aiden Flynn: For me, it starts with trial design. 

One thing I always emphasize is that investing time in study design and planning has an enormous impact on the likelihood of success. Every study is built on assumptions. Sponsors need to understand those assumptions, explore different options, and quantify the risks associated with each path forward. That’s true whether AI is involved or not. 

What AI can do is help expand that process. It can help build the assumption space, analyze historical information, and support risk quantification. Those capabilities can strengthen decision-making before a protocol is finalized. 

Irving Dark: I completely agree. 

One of the most important principles in development is that the best trial design cannot make a bad drug work. But a bad design can absolutely make a good drug fail. We see that repeatedly across the industry. Poor assumptions, underpowered studies, and poorly selected endpoints can undermine otherwise promising programs. 

AI gives us access to more data and greater computational power than we’ve ever had before. That means we can better understand the drivers of success and failure and optimize design decisions earlier. The real value isn’t technology for technology’s sake. It’s using better information to support better decisions. 

There’s a lot of excitement around data. What role does data quality play in AI’s success?

Aiden Flynn: You cannot escape the principle of “rubbish in, rubbish out.” Even the most sophisticated AI model will struggle if it’s built on poor-quality, unrepresentative, or inconsistent data. 

One challenge people often underestimate is how different datasets can be. Electronic health records vary significantly between healthcare systems, countries, organizations, and even hospitals. A solution that works well for one dataset may be completely inappropriate for another. 

That’s why I always advocate starting with the question you’re trying to answer. Once you understand the question, you can determine what data you need and whether those data are fit for purpose. Too often, people start with the technology and then search for a problem to solve. 

Irving Dark: I think that’s exactly right. 

Data interoperability remains a major challenge. Even when you find the right data, there is significant work involved in ingestion, curation of the information, data cleaning, preparing it for analysis, and ensuring that the outputs can support meaningful decisions. The question must always come first. 

Could AI help create a more adaptive approach to clinical trials?

Irving Dark: I believe so. 

Traditionally, development has followed a learn-and-confirm model. You learn in earlier stages and confirm in later stages. What AI may enable is a much more continuous feedback loop throughout a study. 

Imagine systems that continuously ingest data, monitor against design assumptions, identify signals, surface insights, and provide ongoing feedback loop throughout execution. Instead of waiting for major milestones to review information, teams gain earlier visibility into what’s happening in a trial and can make proactive decisions sooner. 

That shifts the focus toward what I often think of as decision science: getting the right information to the people making critical decisions as early as possible. 

Where do you Think AI is Currently Being Overestimated?

Aiden Flynn: Automation is probably the biggest example. 

There’s been a tendency to overpromise what AI can achieve today. While progress has been impressive, many real-world implementations still face challenges with consistency, reliability, and reproducibility. Hallucinations remain an issue, and even the same inputs can sometimes produce different outputs. 

We’re clearly moving in that direction, but I don’t think the technology has fully delivered on some of the boldest promises yet. 

Irving Dark: Accuracy remains a work in progress. 

We’ve all seen examples where AI systems confidently provide answers that are simply incorrect. That doesn’t mean the technology isn’t valuable, but human judgment is still essential. Organizations need governance, oversight, and validation processes to ensure that AI-generated outputs are trustworthy before they influence important decisions. AI enabled systems can be viewed as tools within the developer’s toolkit.  Roles will need to evolve and adapt to utilize these tools in ways that enhance quality while streamlining processes and improving productivity.  

What’s the Biggest Takeaway for Emerging Biotechs? 

Aiden Flynn: Don’t think of AI as the answer. Think of it as a tool. 

The real differentiator is still understanding your development strategy, defining the right questions, and making informed decisions. AI can support that process, but it doesn’t replace the need for expertise. 

Irving Dark: I would add that the future belongs to organizations that combine technology with knowledge-based services. 

As more routine activities become automated, expertise becomes more valuable, not less. The companies that succeed will be those that use AI to enhance scientific, clinical, statistical, and regulatory decision-making rather than simply accelerate existing processes. 

Conclusion

AI has the potential to level parts of the playing field for emerging biotechs, giving smaller teams access to capabilities that were once available only to organizations with extensive resources. Yet technology alone is unlikely to determine success. As AI becomes embedded throughout clinical development, the competitive advantage will increasingly come from knowing how to ask the right questions, interpret the right data, and make the right decisions. 

For small biotechs navigating an increasingly complex development landscape, AI may prove most valuable not as an automation engine, but as a catalyst for smarter, faster, and more confident decision-making.