That scale comes with bold ambition. At EVERSANA, the target is a four-to-one ratio: for every task a human completes, AI agents will carry out four. “Think of one human working with four AI workers,” explains Scott Snyder, the company’s Chief Digital Officer. “We wanted to set a bold target; otherwise, people fall back to incremental thinking.”

It’s an arresting idea – and one that raises as many questions as it answers. Can AI agents really deliver at that level? And how do human colleagues adapt when parts of their roles are absorbed by digital counterparts? The answers, Snyder suggests, lie not just in the technology itself, but in how organisations prepare their people to work alongside it.

For companies under pressure to launch more brands with fewer resources, the temptation to chase every new capability is strong. Executives are pushing for transformation, but organisations often find themselves paralysed by what Snyder describes as the tendency to experiment everywhere without creating the ruthless discipline to pick the cases that are going to drive the most impact.

“Most life sciences companies are stuck right now because the expectations of executives are very high,” he explains. “But what’s happening on the ground is not fast enough.”

One reason is that companies often treat AI as a series of isolated tools, rather than a structural shift. EVERSANA has taken a different tack, testing AI across the full span of commercialisation in ways that mirror how pharma companies themselves operate. “If you don’t take an end-to-end view, you’re going to undershoot the opportunity of AI,” Snyder argues.

At EVERSANA, that rethink has meant breaking down commercial workflows, including campaign strategy, content creation, medical-regulatory review, and dissemination, to name but a few, and rebuilding them with agents in mind. Instead of one person managing each handoff, multiple AI agents can run tasks in parallel, overseen by a human lead. “Why do you do those three steps?” Snyder recalls asking teams. “Why can’t we change that to task 10 agents to go do those things?”

The benefit is not only efficiency, but speed. In a sector where every month shaved off a launch timeline can translate into millions in revenue (and earlier access for patients), the ability to accelerate work is just as valuable as cutting costs. “Time is money for pharma,” Snyder notes. “If I can get to a given market faster with a given product, that’s real economic gain.”

But the transition isn’t automatic. Rebuilding processes means rethinking roles, and making sure the humans in the loop understand when to trust the output – and when to intervene. Agents may be able to carry four times the workload of a single person, but, without the right oversight, that multiplier effect risks creating four times the errors.

The first two are relatively straightforward. ROI ensures use cases deliver business value, rather than novelty. Risk tolerance means knowing where AI’s flaws, such as hallucinations, can be managed, and where they cannot. But the third “R”, readiness, is often overlooked. It speaks less to technology and more to the people using it.

“Training is one thing – that’s the skill part,” Snyder says. “But then there’s the will part, the mindset. If employees aren’t ready to shift the way they work, you can push the best tools and technology on them, and they’re going to reject it.”

That readiness gap explains why many pilots never scale. When AI is framed purely as an efficiency tool – a way to “do more with less” – employees may resist, fearing their jobs will be hollowed out. When it is instead framed as an enabler, a chance to double the number of launches with the same resources, or to focus on higher-value tasks, adoption becomes far easier.

This shift is also giving rise to new roles in pharma’s workforce. Snyder points to translators – employees who understand business processes well enough to identify where AI can add value – and orchestrators, who manage hybrid teams of humans and agents. These roles don’t replace strategists or campaign managers; rather, they reconfigure their work. “Getting those people to be AI savvy so they can connect the dots – that’s gold,” Snyder says.

And while agents take on the heavy lifting, human colleagues remain vital. “The one thing that AI will never replace is the three Cs in a human: creativity, collaboration, and critical thinking,” Snyder insists. “Those are so uniquely human.”

“Real transformation to me is not just, ‘Did I deploy a whole bunch of productivity tools with AI?’” he says. “It’s, ‘Have we changed the way we operate and deliver value to our customers and to the market?’” That shift could eventually redefine business models themselves, with companies moving away from transaction-based metrics and towards outcome-based measures.

In practice, this means new kinds of KPIs. Instead of counting full-time equivalents or project deliverables, Snyder envisions organisations being judged on time-to-therapy, patient adherence, and measurable improvements in health outcomes. “If we’re really doing our job, this is going to show up in patients staying on the regimen and getting healthier,” he explains.

That vision demands more than just deploying agents at scale. It requires designing AI-enabled experiences that help people – employees, healthcare professionals, and patients — grow, rather than simply work faster. Snyder draws a comparison with the early internet: access to endless information didn’t automatically make people better researchers. The same, he suggests, applies to AI. “You just don’t throw AI at people. Designing experiences where they can grow and develop from it, that’s where it’s at.”