This summer marked 50 years since the design of the internet was released in a paper by Vint Cerf and Robert Kahn. That work on a protocol, architecture and philosophy for the sharing of resources between different networks was foundational to today’s world.

Over the following five decades since Cerf and Kahn’s publication we’ve certainly come a long way and there have been a steady stream of major technological milestones.  Those taking us towards the central place online tech occupies today include the first mobile phone in 1983 and 1991’s invention of the World Wide Web, with many more important steps that could be added to that short list.

One of the most recent developments was the public release of Chat-GPT in November 2022 and the subsequent explosion of interest in, and applications for, generative AI (Gen AI). Fuelling this new world has been the proliferation of data – both structured and unstructured – which underpins all Gen AI applications in their quest to speed, simplify, summarise and otherwise support individuals and businesses.

Today, Gen AI is cutting a swathe through almost every sector, including – of course – life sciences and healthcare. What’s particularly noteworthy about Gen AI is the lightning speed with which it’s hit the mainstream, and the last two years have been like no other period in the more than two decades that I’ve been writing about and commentating on ‘digital pharma’.

However, there are many important considerations for commercial pharma executives if they are to benefit from the ascendence of Gen AI in life sciences, starting with the need to assess the current position of this fast-moving technology.

The current state of AI will likely have changed by the time you read these words, and it will certainly keep changing. Today, there’s such a proliferation of data and the use of Gen AI capabilities that many people already view Chat-GPT as old news. Granted, if someone uses just one Gen AI tool then the chances are it will be Chat-GPT, which does have impressive capabilities, but it’s already being overtaken in pharma by specialty Gen AI models such as Microsoft’s BioGPT for biomedical text generation and mining and the drug discovery tools of NVIDIA’s Clara for Biopharma.

‘What’s particularly noteworthy about Gen AI is the lightning speed with which it’s hit the mainstream’

Furthermore, some of the industry’s biggest players are moving beyond off-the-shelf products to develop their own Gen AI models and tools, keenly aware of the need to avoid putting their private data into public tools. Working with firms like Chat-GPT developer OpenAI, a number of pharma companies are building their own ‘walled garden’ Gen AI capabilities so that they – and only they – can leverage all the content and data they have.

For example, there’s the major firm that’s using AI models to augment its content supply chain – including content creation and editing, MLR risk assessment and integrating different analytics sets from different sources. Then there is a midsize company using Gen AI across legal, R&D, manufacturing and commercial, harnessing purpose-built AI assistants to work alongside its employees. Across the industry, examples like this abound.

Such pilots and exploratory initiatives must ensure that the companies behind them are set up to succeed, with the right level of AI readiness. In order to use Gen AI tools to the full extent of their powers there are a number of elements that need to be addressed, starting with understanding what kind of data a company has, how it’s captured and where any gaps exist. From there it’s vital to use the right internal data to keep feeding the algorithms.

The full extent of those powers has yet to be revealed. But for every industry, including pharma, the current focus areas are three-fold. First and foremost, they’re coalescing around efficiency gains and how to leverage artificial intelligence (AI) to facilitate better internal decision-making to drive efficiency of existing processes. In pharma that cuts across everything from research and development to medical to commercial.

The second area that’s being looked at is using AI to build better effectiveness, in terms of better and more impactful customer engagement or content or channel choice in order to define what you should do next. Most commercial pharma organisations are actively looking into, and developing, use cases for those key business areas. The third area is something that a lot of people aren’t talking about… but everybody’s doing, and that’s the behind-the-scenes work.

Initial, early assessments of AI capabilities are widespread, as companies review the cleanliness, appropriateness and completeness of their data, and the quality standards that need to be put in place. In order to ensure that what is fed to the Gen AI engine will help it learn there needs to be a range of process steps to be taken, such as defining what levels of AI understanding are required and how data outcomes will be validated.

Across all three areas one of the most important factors concerns an organisation’s people, in terms of their knowledge and whether they judge AI to have seamless applications within their working lives. Above all it will come down to trust – can AI applications be trusted to be compliant and, as part of that, can an AI tool accurately represent a particular company’s own approach to compliance and risk management? These three steps will serve to position companies for what needs to come next.

For companies to solidify their approach to Gen AI they will need to integrate assessments of their near-term readiness with the three focus areas above. The last two years have seen a tremendous amount of technological change, and the future is perhaps more unknown than it’s been for a long time. But what is clear is that the progress seen with the pilots, often of silo-specific use cases, that have tended to proliferate in pharma to date in Gen AI, and AI more generally, is just the tip of the iceberg.

The next steps will be to understand where a company’s gaps are located, and what benefits could be gained in terms of its digital maturity to deliver customer experience success. Before that’s reached, a firm’s readiness to succeed must be assessed from both a data process and a people perspective, to benchmark the organisation’s maturity and understanding.

‘Above all it will come down to trust – can AI applications be trusted to be compliant’

As the end of 2024 rushes ever closer, with 2025 around the corner, what will the future bring? It’s a difficult question at the best of times, but in Gen AI – where everything’s changing so quickly – there will undoubtably be further developments even before the New Year. That means that, alongside organisational readiness for Gen AI, commercial pharma leaders and senior executives will need to develop their own personal readiness too. They must understand the implications of how this technology is going to change their organisations and what that means for the development of their people and their company. This will require a certain amount of humbleness to address the expertise and knowledge gaps that inevitably exist in such a fast-moving and complex area, while at the same time building the internal capabilities their organisations need.

Furthermore, the current direction of travel is positioning Gen AI to move beyond being just a tool or a system to fix something and starting to look like it could eliminate some of pharma’s functional boundaries. If this does play out it could, for example, enable commercial and medical accountabilities to be handed to the same person – something that would be quite earth-shattering to the way in which pharma currently operates and is organised.

However, while there are ways to future-proof Gen AI, the speed at which the technology is changing means instead that there’s a headlong rush to develop individual models. This sees many firms having multiple activities running through multiple AI engines, with multiple use cases for each, which could be an expensive stage of pharma’s Gen AI journey. But there are a few companies that are taking a smarter path of building AI frameworks that can accommodate different AI engines as needed for certain specialty areas or regulatory codes of practice.

There’s no turning back time on AI, but there are plenty of ways of getting ahead of the curve with the technology’s ongoing evolution.

Dominic Tyer is a Research Director at DT Consulting, an Indegene company