Novasecta’s Principal, Ed Corbett, was asked by The Pharma Letter, in a special report piece, to consider an issue of paramount importance to the future of the industry: AI
Artificial intelligence (AI) has the potential to provide huge benefits to the pharmaceutical industry, from improving R&D productivity through to more effective sales representative deployment and better supply chain management.
Adoption of AI is lagging other sectors, with initial forays mainly being led by big pharma, with its deep pockets and willingness to try new innovations. For many, however, it remains misunderstood, or even feared.
Given the transformative potential of AI, companies must at the very least understand its benefits and develop a strategy that meets each organization’s unique situation. Those that do will be well informed to make decisions; those that don’t, may be left out of the next industrial revolution.
Next industrial revolution?
AI has many definitions and subsets, but at its most basic, it is maths. It uses data that is fed into linked algorithms to create conditional probabilities (as opposed to certainties – a key difference) about desired outputs.
Creating algorithms that improve and refine recommendations autonomously, is machine learning, a subset of AI. Think of the recommendations that you may receive via Netflix, Amazon or iTunes. These combine your preferences (data) with a database of profiles similar to yours and uses an algorithm to make recommendations (probabilities) about things you may like.
The more data the algorithm has, the more accurate the predictions are likely to be. This is something that a human could do, but it would be incredibly labor intensive – and this is one of the principle benefits of AI. It completes tasks that people could do, freeing up time to enable that person to do something a computer cannot. It is for this reason that AI is being seen as the next industrial revolution.
Multiple applications across pharma
Unfortunately, AI also has something of an image problem. Public perception is that it will replace people, leaving many out of work – at its most extreme, AI could lead humans to be controlled by robots. As society becomes more familiar with what AI is and what it can do, many of these fears will be allayed and the technology and its benefits will become more widespread.
For the pharma industry, AI has a huge variety of potential applications – it’s not just one ‘thing’. Companies are applying it to manufacturing to improve packaging, reduce write-offs and limit batch variation. In the commercial sphere, it is being used to guide sales representatives to which doctor or practice is most likely to use a product and refine product messages to drive sales.
It is, however, in R&D that the greatest potential for AI exists. R&D productivity continues to be a challenge, while the cost of developing products continuing to rise. Many companies have turned to M&A activity to plug pipeline gaps, but this does not address one of the fundamental problems with R&D – it is very human-capital intensive and reliant on individual expertise in specific areas for success. AI has the potential to increase R&D productivity by addressing both these challenges.
It can be applied to molecule repurposing, clinical trial recruitment and patient response to drugs, all tasks that are currently highly human intensive. For example, companies like Berg are combining patient biomarkers with AI to better understand the patient response to innovative medicines. This has the potential to not only make clinical trials more efficient and effective, it also will mean that patients in the ‘real world’ get drugs that are more effective for them based on their unique biology.
Varying levels of enthusiasm
Earlier in the drug development pathway, target identification is an area with a huge failure rate and is consequently of great interest to AI-focused companies. Currently, target identification often relies on human expertise to judge whether a molecule may or may not interact with the target site. This requires experience, is often conducted linearly (as often one site is tested at a time) and is consequently labour intensive.
AI can run in silico experiments that test multiple sites and predict what the effect may be. This then narrows down, based on data, the sites that should warrant further human exploration. This is the essence of the benefit of AI – it augments human activity. In fact, the most effective application of AI, is where human input and refinement to the algorithm is a key part of the process.
Pharma companies are approaching AI with varying degrees of enthusiasm. Novartis (NOVN: VX) stands out by putting AI at the heart of its business and many others are making significant, at risk, investments in the technology.
A need for understanding
Given the clear benefits of AI, senior leaders need to at the very least understand what it is, what its potential benefits are and how it may (or may not) fit with the direction of the organization – in short, they need an AI strategy.
The key questions they should answer include which area of the business has most potential for AI, what has been done so far in that area that can be built upon, what inspiration can be gleaned from other companies, what the company prioritizes over the short, medium and long term, and what should the company do to realize the benefits?
For those that see AI as part of the future, a range of partner organizations exist to help companies make it a reality, with most headquartered in the UK and the USA.
However, many of these are unproven – pharma companies need to understand AI in order to cut through the hype and partner with companies that will deliver the intended results.
Fear of human replacement
AI also requires different capabilities and companies need to go ‘to war’ to recruit the best talent. They will also need to change and challenge existing business processes – failure to do this will mean AI benefits will not be realized. Finally, leaders must communicate the benefits of AI to ally pre-existing fears and thereby raise the organization’s ‘AIQ’.
That AI has benefits for the pharma industry is clear. Whether it is right for each company depends on each individual organization, but leaders need to understand potential benefits and develop a strategy for their unique situation. An absence of a strategy may lead to awkward questions from shareholders and leaders may not be replaced by a robot, but by another human.
To view the article in The Pharma Letter, click here.