By: Joshua Finley

The increasing use of generative AI and the continued evolution of the technology have been the subject of many news reports. Users have also embraced the technology, for instance, OpenAI’s ChatGPT was launched in November 2022 and attracted 1 million users within a number of days of operation. In just months the chatbot had 100 million monthly active users.

Enterprise AI adoption has surged as companies seek to harness its ability to quickly generate insights from vast data sets. Consulting firms have eagerly embraced generative AI’s potential to transform 21st-century business. In June 2023, McKinsey projected that generative AI could deliver between $2.6 trillion and $4.4 trillion in annual economic benefits across industries. However, according to IBM’s Institute for Business Value, 60% of organizations are not yet developing a consistent, enterprise-wide approach to generative AI.

Back to Reality

The sobering reality is that AI contribution to economies across the globe has not reached these giddy heights, at least not yet. One person who can shed some light on the stuttering adoption of AI is IBM’s Julien Willard.

Dr. Willard is a strategy leader for life sciences at IBM Consulting, an advisory board member at Harvard Business Review, and previously led digital and AI strategy across insurance, health and life sciences sectors at Accenture. Prior to his strategy consulting work, he served as a diplomat and health economist on the issues of market access, health economics and outcomes research, and real world evidence. He is deeply immersed in AI and big data, from strategy to execution.

According to Willard, the growth of AI applications has not met the expectations set by many industry analysts, due to several factors. First, there is the uncertainty surrounding growth. Second, concerns about data privacy continue to hinder progress across industries, and in pharmaceuticals and biotech in particular. Third, the usability of enterprise data remains a significant challenge, as many organizations struggle with fragmented, incomplete, or poorly governed data sets.

“Enterprise IT spend is around 4.7 trillion dollars and about forty billion is expenditures related to enterprise AI. Drill down further and you find that spending on generative AI is only a drop in the bucket, close to 2.5 billion. Business leaders are playing the waiting game to see if AI can prove that it’s a technology that can provide growth which can total millions of dollars.”

“Secondly, there’s the issue of data, especially in the life sciences sector. Many clients that I serve are cautious in their approach to AI adoption, and many are still wrestling with educating internal stakeholders that technology solutions today can mitigate any potential patient privacy concerns and ensure data security,” says Willard. On February 21, 2024, a ransomware attack on Change Healthcare exposed protected health information affecting up to one-third of Americans.

Willard highlights another challenge with data, particularly in the life sciences sector, where it is often siloed within companies and across the value chain. This fragmentation makes it difficult to organize data in a way that enables effective AI adoption and hampers efforts to derive meaningful insights.

“Big pharma companies have disparate datasets, seated in different parts of the organization and available in various formats, for instance, they may have disparate data coming in from the field through clinical trials, remote monitoring devices and electronic health records. That all must be standardized before it becomes usable for any machine learning techniques,” Julien Willard continues.

“There’s an overused industry analogy comparing data to oil. However, if unrefined and systematized, it cannot really be used. Many AI investment decisions today are being held back due to concerns around data quality, governance, and effective data management. You can see a similar trend in the startup and investor community, where many investments are being delayed or put under conditions until enterprises have their data properly organized and managed.”

The Focus Challenge 

Separately, large-scale AI implementation is often hampered by what Willard calls ‘The Focus Challenge.’ Instead of prioritizing a few high-value use cases that directly impact the bottom or peak line, companies spread their efforts too thin across multiple initiatives, preventing them from scaling effectively. Additionally, existing staff may lack the full technical expertise needed for advanced AI use cases, and with their daily responsibilities, proof-of-concepts become an added burden, further diluting focus and progress. “There’s also the challenge of whether the company’s operating model is geared to support the scale of the AI implementation,” says Julien Willard.

The Silver Lining

For Willard, AI is already driving significant advancements in the pharmaceutical industry. Specific use cases include prioritizing indications as part of asset strategy, identifying and enrolling patients for clinical trials, delivering personalized content and outreach to patients, and streamlining processes like medical writing. These applications help accelerate development, improve patient engagement, and enhance overall efficiency across the value chain.

“COVID-19 actually accelerated  AI adoption in drug development, and today we have a few drug candidates in clinical trials that were developed through an AI-assisted drug discovery process. While a moment of crisis, the pandemic represented a huge learning moment for the industry, and increased the pace of innovation that had been taking place for decades.”

Generative AI solutions are also being designed to manage repetitive tasks, with many poised to impact consumers directly. 

“No one wants to be on the phone waiting to speak with HR to get basic paperwork for a mortgage application, for example. Clients want to use intelligent chatbot capabilities to automate this process and let employees handle most of the tasks in an automated fashion. This relieves the HR team to do value-added work, focusing on talent development,” says Willard.

The same principle applies to case report forms used by pharmaceutical companies to gather patient information in clinical trials. After seeing a patient, clinical trial personnel must fill out a form with hundreds of fields. Currently, each form is manually designed for specific trials, making the process time-consuming and labor-intensive. Once the forms are collected, they tend to follow a manual analysis and review. Intelligent automation is now becoming table stakes.

“This task is perfect for AI. Companies have done it manually for years, but now AI can pull templates from historic forms and use them as a baseline, generating unique outputs based on the uploaded clinical protocol. We’ve implemented this model and reached about a high percentage completion of the forms.”

The remaining percentage is where human and human intelligence step in. “This is another reason why those predicting AI will replace humans are wrong, at least in the medium term,” says Willard.

The Future

According to Willard, AI isn’t a Pandora’s Box; it won’t lead to mass job losses. AI and human beings are an opportunity for collaboration. Evidence suggests that AI can transform how we manage data and how businesses and consumers alike can benefit from the technology.

Julien Willard currently applies his expertise to solving complex challenges for executive teams at leading pharmaceutical and biotech companies. For insights into the fast-evolving AI landscape and its effects on healthcare and the life sciences industry, visit his LinkedIn page or personal website.

Published by: Khy Talara