Skip to main content Skip to footer
Cognizant logo


October 27, 2023

4 ways clinical development will improve with gen AI

By changing their approach to clinical development with generative AI, life sciences companies can realize exponential revenue opportunities and offer renewed hope to patients.


When it comes to clinical development, saving time translates to saving lives, or at least improving them, through faster availability of treatments. For life sciences companies, it also translates into revenue opportunity—and a significant one at that. Some industry sources estimate that delivering new treatments to market ahead of schedule can be worth anywhere from $600,000 to $8 million per day.

The promise of speed is why generative AI is so compelling for life sciences companies. Generative AI helps automate standard development processes, find patterns, identify insights and produce relevant, meaningful content in any number of formats, all in a fraction of the time it would take for a human to do so.

We’ve identified four key areas of the clinical development lifecycle that can be accelerated and enhanced through the use of generative AI and supporting technologies. For life sciences companies that adopt gen AI, clinical development efficiency gains will be measured not in days or weeks but in months or even years—representing both an exponential revenue opportunity and renewed hope for patients.

Four ways to use gen AI in clinical development

#1: Streamline research processes using semantic search. Generative AI could fundamentally change the clinical and scientific research process. Instead of beginning with a manual keyword search and sifting through hundreds of articles across various sources, research teams could prompt a generative AI-enabled tool to rapidly search, gather and distill relevant articles or even suggest unanticipated information pathways to explore. This frees researchers to focus on analysis.

In our client engagements, we’ve seen 10X time savings when life sciences companies have applied generative AI to the research process.

Generative AI could also change how research is conducted. Because the underlying models understand intent and context, these tools can work from a mission- or goal-based prompt, as opposed to relying on traditional keyword searches.

For example, if a company is developing a cholesterol medication, a traditional process might begin with keyword searches involving different combinations of terms like "cholesterol," "clinical trial,” "efficacy," "safety," "hyperlipidemia" and "statins," ultimately uncovering disparate results across different source platforms.

However, with an AI-enabled tool, researchers could state their goal and receive contextualized reference materials to support that specific concept. “I'm designing a clinical trial for a new cholesterol medication,” a researcher might state to a generative AI tool. “Please provide an overview of the most recent studies on cholesterol medications, their efficacy and safety, and how they compare with existing treatments. Also, any insights on best practices and regulations would be valuable."

Finally, generative AI can also be used to analyze and synthesize relevant research materials and present them in a digestible way to a variety of constituents, from team members across multiple departments, to regulatory agencies, to institutional and ethical review boards. Ideally, these summaries would include a high degree of explainability regarding why those materials were selected. This further expedites the review process, which can also unlock time savings.

Explainability is the ability of the generative AI-enabled tool to document how it arrived at the conclusions presented in a given output and provide a rationale for any recommendations being proposed.

For example, when authoring a clinical trial protocol, the tool may include references to past protocols in a government database with similar trial dynamics, as well as components from successful trials.

This capability can help researchers and other team members quickly assess the output, view it within the proper context and identify the best next step in the process.


#2: Automate the authoring of clinical trial protocols to improve efficiency and access.
It can take a few months to over a year to author a clinical trial protocol document. But companies that leverage generative AI and large language models (LLMs) can dramatically reduce that time, to days or even hours.

A foundational model for research can be fed and trained on thousands of clinical trial protocols available online through government and industry databases and proprietary company research data. As such, generative AI-enabled tools can quickly scan thousands of entries within a given database and identify the patterns relevant to investigational products, certain conditions, specific patient populations or other factors. If the company has its own proprietary data, the tool can also incorporate that information into the foundational research model.

As the tool identifies relevant patterns, it can design a baseline study, define the narrative, determine eligibility, draft exclusionary criteria and provide other details. The draft, or a group of draft options, can then be evaluated and refined by a human.

It’s also possible to expedite the protocol review and approval cycles because the protocol could be summarized and accompanied by an explanation. We estimate that life sciences companies can shave 50% to 60% off the time spent writing protocols, while reducing related review cycles another 10% to 20%.

50%-60% estimated reduction in clinical trial authoring times.


Life sciences companies can also use generative AI to design studies that leverage other digital capabilities or enable a non-traditional trial model, such as a decentralized, virtual or hybrid format.

For example, during the authoring phase, the foundational model or LLM could leverage valid data from wearables, sensors or other means of collecting biomarkers outside of a traditional clinical trial site or clinical setting. When they have specific information about the technologies needed to enable such a trial, companies can consider a wider range of options, which may, in turn, allow them to recruit from a larger base of participants.

In this way, generative AI isn’t just expediting the authoring process, but potentially designing a trial format that would speed recruitment, enrollment and testing.

Beyond identifying eligible trial participants, generative AI could also analyze demographic and behavioral data to suggest population segments likely to use these tools and remain engaged throughout the study.

One high-profile example of how generative AI is being used within the medical research community is AlphaFold. This AI program, developed by Google DeepMind, leverages deep learning to perform predictions of protein structures based on its string of amino acids. With this technology, what once took years of research can now be reduced to minutes.


#3: Identify areas for improvement through next-gen clinical study reports. It’s very time-consuming to analyze clinical trial data and summarize it in a way that demonstrates statistical significance, while also maintaining safety and efficacy standards. It’s also prone to error when done manually.

For years, conventional AI has been used to accelerate this compilation of data through prediction and probability. However, generative AI can reveal new connections not immediately apparent to human researchers or conventional models.

For example, traditional AI may be able to extract a key statistic, such as trial recruitment velocity. However, with generative AI, the model could explain the reason behind that number and also offer recommendations for improvement.

The additional context and recommendations could include:

  • They are behind schedule by 40% and need to recruit another 120 people in two months to meet the current deadline

  • Social media channels, like Facebook and Instagram, are the highest-performing recruitment channels

  • Increasing the social media promotion budget by 60% (and reducing traditional ad buys) is likely to help them meet their recruitment goals. The gen AI tool could generate content relevant to specific trial and patient populations, including messaging, visuals and tone for these promotions

  • Further refining the target audience for social ads could accelerate enrollment by up to 10%

One of the major value drivers of generative AI is that this information is all conveyed in an easy-to-understand format. No longer would research teams need to review several pages of Excel-based data or perform manual calculations to see how changes in the strategy may affect outcomes. Nor would they need to cross-reference data with social media plans or ad buys. Instead, the tool takes all that information and consolidates it into a clear list of recommended actions.

Yseop, an AI software company working with enterprises in highly regulated industries, developed Yseop Copilot, a secure content automation solution. Yseop Copilot leverages pre-trained LLM models for the biopharma industry to automate scientific content creation, helping medical writers across hundreds of clinical trials to significantly cut writing times.


#4: Expedite the treatment launch process.
Once approval for a new therapy is secured in a primary market, a tremendous amount of work goes into launching the drug in a secondary market. This includes strategy development, market research, contracting, pharmacy benefit manager (PBM) negotiations, agency engagement, content creation and material development. As with the research and protocol writing processes, much of this activity could be automated through generative AI. And, as noted above, gen AI-enabled tools can also provide recommended actions to improve outcomes or further accelerate the process.

We estimate that companies could generate another 10X time savings, compared with traditional processes, through the use of gen AI in this area. For example, when a drug is close to gaining approval, generative AI can help commercial teams research and compile baseline strategy documents for secondary markets, taking into account any specific regulations that need to be adhered to.

In addition, generative AI-enabled tools can be used to translate existing materials, such as ads, websites, brochures or other sales materials, so that these assets reflect the local language and culture. Using the technology in this way could shed an additional year from the go-to-market timeline and dramatically reduce costs related to content creation and design across various markets.

Adopting and scaling gen AI in clinical development

While generative AI will undoubtedly play a role in the future of the life sciences industry, factors such as an uncertain regulatory landscape, coupled with the rapidly evolving nature of generative AI itself, are prompting some companies to delay investment until the course ahead is clearer.

However, those who take a strategic approach at this early juncture and build maturity will likely find themselves with a strong competitive advantage as the rest of the industry plays catchup.

With that in mind, here are six recommendations for building generative AI maturity today:

  1. Foster a culture of AI literacy

    To reap the full value of generative AI, everyone in the business needs a basic understanding of how generative AI will reshape their functional area. That learning process should begin now, through training courses that focus on the specific opportunities and benefits generative AI presents for each team, as well as its potential to disrupt some existing workflows. Ultimately, the energy and excitement of people will help improve the speed of adoption, maturity and scaling of more advanced use cases.

  2. Build a robust partnership ecosystem

    With the scarcity of generative AI expertise, companies need to build an ecosystem of partners, including relationships with academic institutions, hyper-scalers, data providers and specialty gen AI vendors to accelerate knowledge building and implementation of gen AI solutions. At the same time, it is important to work with a business transformation partner that can help the organization develop its internal capabilities over time. In the case of life sciences, it is critical to select an organization that not only understands the technology but also the industry, its challenges and regulatory landscape.

  3. Establish governance

    Ethical, responsible and safe use of generative AI is especially important for life sciences companies, which need to guard both sensitive patient data and intellectual property. As regional and industry standards develop, it is up to organizations to assess and mitigate risk and biases consistently and actively. This is another area where a trusted partner can help, assisting organizations as they develop, deploy and oversee necessary protocols to ensure the technology is being used safely and responsibly.

  4. Experiment and learn

    Now is the time to empower teams to explore use cases, experiment with the technology and run small-scale pilots to further those efforts and identify high-value opportunities to scale. To support our clients in this journey, we have developed a use case prioritization framework and other tools to help companies identify high-value use cases and evolve ideas into proofs of concept and production solutions.

  5. Own critical expertise

    It’s critical to establish a center of excellence (CoE) or other body to oversee the strategy and activity for the organization as a whole, as well as manage staff upskilling and development. This CoE should also identify best practices and develop frameworks and repeatable solution accelerators to institutionalize key capabilities.

  6. Plan for adaptability

    Because the technology is still evolving, as is the regulatory landscape, companies should take an agile approach so they can adapt their strategies as generative AI matures. In addition, companies should align audit and regulatory activity to support explainability and transparency and enable rapid adaptation if and when the landscape requires it. Designing flexible strategies and initiatives is the key to success in a high-potential but uncertain environment.

Unlocking gen AI impact in clinical development

Most of the use cases explored here are relatively straightforward, both in terms of their practicality and low barriers to deployment. They represent a low-risk, high-reward value proposition.

However, far more advanced use cases represent the next generation of this technology, such as using generative AI-enabled telehealth tools for patient companion apps. These applications, which would mimic a traditional caregiver experience and provide one-on-one patient support, present a tremendous opportunity to improve access to care, while also addressing significant shortages of healthcare professionals.

As life sciences companies experiment with generative AI, they will begin to lay the foundation needed to harness its full potential for both their own business and the people they serve.

To learn more, visit the Generative AI and Life Sciences sections of our website or contact us.



Bryan Hill
Life Sciences Chief Technology Officer
Bryan Hill

Bryan Hill is Chief Technology Officer for Cognizant Life Sciences, responsible for digital solutions and technology innovation. His focus is on how emerging tech can help clients increase innovation to bring new therapies to market faster.

Bryan.Hill@cognizant.com



Latest posts

Related posts

Subscribe for more and stay relevant

The Modern Business newsletter delivers monthly insights to help your business adapt, evolve, and respond—as if on intuition