AI may be a buzzword, but unless healthcare and life sciences organisations think carefully about strategy, data and capability, it rarely delivers real productivity gains.
When Henry Ford introduced the production line and the Model T, it changed the world. Affordable mass-produced cars transformed how people travelled and reshaped cities and countries.
But there was a problem. The infrastructure to support this change did not yet exist. Roads were not designed for large numbers of cars. There were no road rules, no driving licences and no vehicle registration systems.
AI is at a similar moment today, says Daniel Bacon, a Principal at Veeva Business Consulting APAC. It can and will transform every aspect of our working lives, but like the era of the first mass-produced cars, while the technology is rapidly progressing, the infrastructure is not there yet; from the working systems needed to use it properly and realise productivity gains to the training and upskilling in critical thinking needed to avoid the seduction of fast, easy, but misleading AI outputs.
Still, the opportunity is vast. For businesses, it’s a critical tool that will help them evolve to a Human by Exception model.
“There is a vast opportunity to transform our businesses, increase efficiencies and reduce the human workload of routine low value tasks, instead focusing critical human resources on high complexity, high impact activities.” Daniel says.
“However, the common misconception is that AI is a mature technology, and that it’s an off the shelf solution.”
Why AI doesn’t automatically equal efficiency
Research conducted by MIT last year uncovered a disturbing truth: just 5% of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable P&L impact. Sixty percent of organizations evaluated enterprise-level or custom tools, but only 20 percent reached pilot stage and just 5 percent reached production. Most fail due to brittle workflows, lack of contextual learning, and misalignment with day-to-day operations.
MIT’s research scored the healthcare and pharma industry just 0.5 out of a possible 5 for Gen AI disruption, on factors include emergence of new AI-driven business models and changes in user behaviour attributable to GenAI, including cognitive off loading and automation bias, even by users considered experts in their fields.
For Daniel, the challenge for organisations is therefore not simply adopting AI tools, but building the broader system and culture required to use them effectively. He also cautions that this is not something leaders can assume is a future issue. While organisations are grappling with how best to implement their strategy, their employees already have access to enterprise & personal AI tools in their day-to-day work – what he calls the ‘rise of the shadow AI ecosystem’.
“Individual staff members are using tools like Microsoft’s Copilot. Because it comes automatically with their Microsoft tools, all of a sudden, they have a generative AI solution asking to help with their emails.”
For someone looking to get through their email backlog, or respond rapidly to a request from their boss or a customer, it becomes hard to resist what former politician and Chief Executive Officer @ Future Government Institute Victor Dominello referred to as ‘The Seduction of the Yes Button’.. He recommends AI adoption with ‘choice architecture’ – in other words, deliberate reflection before adoption.
This underlines that it’s one thing to have the tools, but another to have the right data, capabilities and strategy to realise real gains.
“We want to use AI, but if it is not implemented intentionally, with clear KPIs and productivity objectives, it is likely to only achieve an increase in your IT costs without measurable productivity and quality benefits.” Daniel says.
(source: MIT)
Get your strategy right
The challenge then becomes leaders using AI to design more effective workforce systems rather than just automate tasks.
Daniel points to an industry example of how this can work: Sanofi’s intentional approach. He recently interviewed Liz Selby, the GM of Sanofi Australia and Lisa Currie their head of Transformation, at Veeva’s Australian Commercial Summit and the company has positioned itself as an ‘AI-powered biopharma company’, deploying AI across research, manufacturing and commercial operations as part of a broader digital transformation strategy.
The goal, Daniel says, is not simply to automate tasks but to redesign workflows and utilise what is known as agentic AI to improve both efficiency and quality.
In biopharma, Daniel points to one example where AI can significantly reduce commercial & medical content review time as well as quality: Veeva’s ‘Quick Check’ & ‘Content” agents. These agentic AI tools can help cut the review process for documents from 17 to 21 days down to a matter of hours, and with far less room for error.
The key is to bring it back to your goal in introducing AI into the business.”
“It’s easy to say you want productivity and efficiency, but that’s too high level. Be specific about the areas that are complex, routine and high value, and understand how these augment or replace the critical roles and functions in your business. This where AI can make a real difference,” Daniel explains.
“The aim isn’t just to make processes faster, but to improve the quality as well. That’s one of the clearest ways to distinguish thoughtful AI integration from shallow automation.”
Value-chain mapping can be a real asset here, as it helps work out where AI can remove friction in business processes, rather than just speeding it up.
Then once you’ve decided where in the business to prioritise, the next stage is to work out planned outcomes, and importantly, how to measure the gains from the AI solution.
He also notes that while tools like ChatGPT and Gemini have made AI widely accessible, organisations should consider whether more tailored solutions are needed for their specific workflows.
Focus on implementation
Once the strategy is clear, the next challenge is implementation. That starts with the fundamentals. AI tools need high-quality data, context and workflows that have been deliberately redesigned around how the technology will be used in practice.
“AI agents are designed to work within a workflow. If you don’t understand the workflow you get sub optimal outcomes.”
Businesses must map the workflows then redesign those same workflows to take advantage of the AI solution being implemented
In practice, Daniel says it helps to think across two-time horizons. In the short term, we should be prioritising an AI-augmented human model, where people use AI to improve the quality and speed of their work rather than hand over entire processes.
How this works in practice can be seen in professional chess. After computers began defeating grandmasters, five time World Chess Champion Magnus Carlsen and others started experimenting with what then became known as ‘centaur chess’, where human players use computer engines to analyse positions but still make the final decisions. These human–AI teams frequently outperform both top human players and standalone chess engines, demonstrating the power of combining human judgement with machine calculation.
“In this scenario Human plus AI beats AI alone,” Daniel says.
That shift also gives leaders a practical way to measure value. Rather than asking whether AI is being used at all, the better question is where it is being embedded into workflows, and what measurable difference it is making. The simplest measure here is speed – how long task taking compared with what they used to.
Training teams and building capability
For leaders, the challenge is deciding which capabilities to prioritise in their teams, so AI adoption translates into real productivity gains rather than isolated experimentation. Daniel highlights a couple of key areas.
The first is effective prompt engineering and iteration; helping teams move beyond basic prompts to managing more complex AI workflows.
“That’s building people’s capabilities and equipping them to use AI more effectively, it also helps avoid the automation bias, by building an understanding of how to use AI, not just see it as an automation tool.”
In practice, this means teaching staff how to refine prompts, interpret outputs and iterate on results, rather than treating AI responses as final answers. He explains that because generative AI still depends heavily on the quality of the prompt, the context it is given and the human judgement applied to its outputs.
“We expect AI solutions to be perfect. But generative AI solutions are probabilistic and limited by the data they have access to, which means, you shouldn’t just take the first answer. Treat it as your 80 percent first draft.”
The second capability is what Daniel calls data empathy – helping non-technical managers understand how data influences decisions.
“For a long time, we used to talk about data driven decision making. The issue of a data driven decision making paradigm is it dehumanizes the process. If we try to bring humans and AI together then we introduce a term called Data empathy.”
“This is understanding how the data is influencing the human decision making that we’re doing.”
In other words, leaders need to help teams understand not just what the data says, but how it shapes judgement and decision-making.
Often, Daniel says, organisations will find those shadow AI staff – or those already experimenting with AI – are the likely early adopters.
“The people that come on the journey fastest are the people who are already trying to do it themselves,” he explains.
Human by exception
Longer term, the impact of AI on organisational design is likely to be significant, with new roles emerging as companies adapt their operating models.
“There will be changes to organisational structure. There will be new functions in the business,” Daniel explains.
Many of these roles reflect how AI systems actually work. While large language models sit at the core, they rely on context, governance and constantly updated information to produce reliable outputs. Context engineers will help maintain the data, workflows and market insights that feed into these systems, ensuring the AI has the right information to support decision-making.
In regulated industries such as healthcare and life sciences, compliance architects will play an equally important role, ensuring that evolving regulations and governance requirements are built into these systems from the outset.
At the same time, human–AI designers will focus on how people interact with AI tools, ensuring that systems are intuitive and that the interface between humans and machines supports better decision-making.
As these systems become embedded in everyday workflows, Daniel says organisations will move toward what he calls a “human by exception” model.
In this approach, routine interactions and operational tasks are handled by automated systems, with people stepping in when judgement or escalation is required.
“It’s your escalation process. It’s similar to how chat bots work manage routine, frequently asked questions but you’re eventually given the option to talk to a human being for more complex or ambiguous issues.”
However, he emphasises that some situations will always require human judgement.
“You need to send a human being in to negotiate that contract. You need a human being to intervene with an underperforming employee or a dissatisfied customer.”
Ultimately, he says, organisations still operate in human markets.
“Most business are selling to humans and providing services for human beings. We generally don’t sell or deal with another AI.”
Actions to take now
For leaders in healthcare and life sciences organisations, this shift raises an immediate strategic question: what are the things they should be doing right now? One of the first priorities is ensuring the right parts of the organisation are working closely together to design and implement AI strategies.
“Leadership teams need to look at how closely they are bringing their L&D, their commercial and their IT teams together,” Daniel says.
In the near term, learning and development teams will play a critical role as organisations move into what Daniel describes as an AI-augmented human phase.
“They need to take a long-term approach. This is not just a short-term capability drive,” he adds.
Over time, leadership roles themselves will also evolve.
“Managers’ roles are going to change from managing people and tasks to curating knowledge and curating processes.”
In the shorter term, however, leaders will still need to focus on supporting their teams through the transition.
“Managers don’t just need to communicate, they also need to coach. They also need to support, because people will hit walls and need emotional engagement.”
For leaders, the challenge is not deciding whether AI will transform organisations, but how intentionally they design that transformation.
Like the early days of the mass-produced automobile, the technology has arrived before the infrastructure around it. The organisations that benefit most will be those that build the systems, data foundations and capabilities needed to use it effectively. Those that do will unlock what Daniel describes as the real promise of AI: the ability to grow the topline without growing staff, improve decision-making and redesign work without simply adding more people.
In other words, the opportunity is about more than automation. It is building organisations where humans and AI work together, each focused on what they do best.
About Hunton Executive
Hunton Executive is a specialist executive search firm dedicated exclusively to healthcare and life sciences. We work across the full healthcare and life sciences ecosystem – from early-stage innovation through to multinational scale and healthcare delivery – supporting organisations at the moments where leadership decisions shape growth, performance and long-term value.
Discover how Hunton Executive can help you secure the leadership talent that drives real impact – strategically, seamlessly, and with a true partner mindset.
If you’re planning your next executive hire or want to strengthen your leadership pipeline, let’s start the conversation. Schedule a call >>