As per a new Axendia Market Research Report: The State of Generative AI in Life Sciences: The Good, the Bad, and the Ugly, among the pharmaceutical DX-digital transformation pros, 67% went on to answer- not ready when asked about how ready the life sciences sector is to effectively go ahead and leverage generative AI when it comes to infrastructure, regulatory considerations, and expertise. When it came to R&D professionals, there were 43% who said the same.
Almost 50% of all the 200 survey respondents anticipated that it was indeed going to be over 4 years before generative AI could go on to become mainstream in their functional area.
The early use cases in terms of AI in the biopharma sector, or at least the early public use cases, have come from R&D, and that too more specifically from target identification as well as molecular design. The real questions that need to be asked are: beyond discoveries and designs, where are the use cases when it comes to biologics manufacturing, QMS, supply chain management, as well as operations, and what is holding one back.
AI when it comes to biopharmaceutical manufacturing
A massive 79% of the respondents in the study said that they believed generative AI happens to have the potential to bring about a revolution in the drug manufacturing landscape in terms of quality as well as efficiency. However, the percentage of those respondents who are claiming to be present generative AI users- 5% goes on to paint a picture that’s evolutionary and not revolutionary.
The application fruit, which is indeed low hanging as per the industry, is kind of a far-fetched process modelling and simulation in which over a quarter of the respondents said that generative AI happens to be having the potential to be the most beneficial of all. It also happens to be an area wherein most of the drug manufacturers are flush with a resource that is requisite to AI, and that’s data.
It is well to be noted that process optimization at 19%, drug synthesis as well as formulation at 12%, and real-time tracking at 10% are the top four opportunities that happen to be the most cited.
Apart from the data access, there is one more reason why modeling and simulation happen to be seen as the first and even the best opportunity when it comes to generative AI in order to make an impact on pharma. It is the discipline that is least encumbered by external forces, which create a pause when a change to formulation, process, or tracking is considered. The fact is that regulatory influence happens to be primary amongst such forces, which goes without saying. However, the numbers do reinforce the truth, which is that 69% of the respondents to the Axendria survey went on to say that regulatory compliance concerns happen to be the top challenge or barrier in terms of executing generative AI within the gamut of drug manufacturing. Data security, as well as the workforce’s knowledge or lack thereof, happen to be tied for a distant second spot in terms of the challenge or barrier category at 36%. It indeed looks like there is a massive chasm between the workforce and C-suite when it comes to knowledge. When asked how familiar they are with the generative AI concept, 82% of the C-suite executives remarked that they were very familiar, but their confidence waned as the question percolated through the ranks, with 62% of the VPs and GMs and 31% of individual contributors. Only 23% of the heads of the department as well as the directors said the same.
AI in pharma supply chain management
Still rattled due to the pandemic supply chain effect, a third of the respondents to the report from Axendia Market Research said that they are looking out for AI for a leg up so as to avoid shortfalls in supply chain management in the future. Be it demand forecasting, predictive analysis, or inventory management, all happen to be the top applications that expect generative AI to help. While there were 56% of the respondents who said that they are either confident or very confident that generative AI can very well upgrade the efficiency and resilience of the supply chains, around 77% said that they are not at present making use of AI-driven technology or other analytical tools so as to manage the supply chain resilience. So who is to blame for such a dearth of effort?
For starters, it is the supply chain disparity. It is well to be noted that in retail, it is the retailers that rule. Walmart, apparently has the selling power to force all of its suppliers to either do it their way or go to Amazon. This kind of power makes data aggregation as well as the ensuing analysis pretty seamless. Beyond the select few who are at the top of the biopharma pyramid, no other power exists within the life sciences.
AI when it comes to QMS
Post-market surveillance as well as QMS happened to rally to high levels in terms of support for generative AI solutions within the Axendia report, with 77% going ahead and indicating that performance reporting as well as metrics enhancement, as well as efficiencies are going to hold a prominent potential. However, the present adoption of technology for that particular application is pretty low, as it stands at 12%.
This happens to be yet another case of an application for tech in a data-rich discipline; however, it is the quality and governance of that particular data that give the would-be adopters pause. It is well to be noted that a massive 88% of the respondents to the survey remarked that they think generative AI can go on to introduce some potential quality risks.
AI when it comes to lab operations
Notably, data analysis in terms of lab operations has gone on to garner the most present adoption when it comes to generative AI technology. Only 2 in 10 survey respondents said that they happen to be using it in spite of 70%, indicating that they believe the technology happens to hold the potential to go ahead and revolutionize the lab process when it comes to efficiency as well as quality. It is the data analysis as well as interpretation that happens to lead the charge, with 94% stating those exercises would benefit most from the AI, which is distantly followed by workflow as well as process optimization at 54%.
