Closing the Medical Writing Capacity Gap: Why Training Beats Technology

When Applied Clinical Trials looked at resourcing and outsourcing trends in drug development last year, the pattern was clear: biopharma leaders are moving away from a single full-service CRO relationship and toward hybrid models that blend in-house expertise, trusted consultants, and targeted outsourcing. As leaders of functional service provider (FSP) medical writing teams, we've seen a similar shift in medical writing, though it largely reflects capacity demands rather than a change in resourcing strategy. More importantly, it has revealed a challenge that hybrid contracting alone cannot solve: a shrinking talent pool with increasingly uneven levels of experience.

Medical writing teams are expected to deliver a higher volume at greater speed while meeting growing demands for specialized expertise, driven in part by the explosion of activity in obesity and metabolic disease that is compressing timelines across the industry. Artificial intelligence (AI) is often touted as the solution. Although it has a role to play and can get sponsors part of the way there, it cannot do the one thing this momentrequires: develop the next generation of writers who know how to use it well. Based on our experience, we believe this challenge is fundamentally about talent development and training, not technology.

Why hybrid models are winning

Historically, sponsors have swung between staff augmentation and full-service outsourcing. Today, most are landing somewhere in between. Rather than handing an entire program to one CRO or building an all-FSP bench from scratch, sponsors are combining dedicated resources with deliverable-based support, and adjusting the mix as programs evolve. A 2024 industry survey found 41% of sponsors had recently increased their FSP outsourcing, compared with just 27% who increased full-service outsourcing, and that shift shows no sign of reversing.¹

The appeal is not just flexibility, but the ability to scale resources while remaining closely connected to the work, the team, and the long-term program strategy. A hybrid model lets sponsors expand writing capacity during periods of high demand and reduce it when needs stabilize, while preserving continuity on programs they have spent years advancing. However, that advantage depends on access to experienced talent, bringing the industry’s talent challenge back to the forefront.

The post-COVID experience gap

Since 2020, experienced medical writers have become harder to find and more expensive to recruit. Some of that is broader life-sciences hiring pressure: clinical and regulatory roles across the industry face a limited pipeline, and competitors actively recruit writers the moment they become available. We have observed that senior writers are staying put rather than changing employers, likely reflecting broader uncertainty in the job market. The result is a tighter talent market, with fewer opportunities for organizations to recruit experienced writers externally.

At the same time, the industry has seen an influx of new writers, which bodes well for the future talent pipeline. However, it has also exposed a widening gap between early career and highly experienced talent. Organizations increasingly face a choice between less experienced writers who can add capacity but require substantial guidance, and a limited pool of veteran writers who are already stretched thin. Neither end of the spectrum, on its own, is a viable long-term capacity strategy.

Where AI helps, and where it still falls short

When used effectively, AI can streamline repetitive tasks, accelerate information retrieval, and free writers to focus on higher-value work requiring contextual understanding, strategic thinking, and expert judgment. Those benefits matter for teams operating under growing capacity pressures, even if they do not address the underlying talent challenge.

However, the industry’s enthusiasm for AI could outpace the realities of responsible implementation. Implementing AI at scale requires substantial investment in governance, validation, security, and oversight. For vendors supporting multiple sponsors, data privacy and system segregation requirements add significant complexity. Quality also remains a concern. Studies of large language models in clinical and regulatory settings have reported hallucination rates ranging from 8% to 20%, making expert human review indispensable. In practice, an AI can augment experienced writers, but it cannot replace the expertise required to assess accuracy, apply context, and exercise judgment.²

That reality raises a broader question for the profession: if AI takes on more of the first-draft process, how will the next generation of writers develop the judgment required to review it effectively? Scientific interpretation, regulatory risk assessment, and the ability to recognize when something deserves a second look are capabilities built through experience. If early career writers spend their formative years refining AI-generated content rather than creating and defending their own work, they may never fully develop those skills. AI still has a role to play. The point is to deploy it deliberately, treating critical thinking, contextual understanding, and professional judgment as capabilities that must be actively developed.

Training and mentorship as strategy, not an HR afterthought

Training and mentorship offer the most durable response to the capacity challenge. In our experience, if experienced talent is becoming harder to hire, organizations must develop it internally, and that requires more than an onboarding checklist.

Effective development starts with creating clear pathways to expertise. Newer writers need access to experienced colleagues who can help them navigate the nuances of different document types, therapeutic areas, and regulatory expectations. The strongest trainers tend to be the subject matter experts themselves, who understand not only the guidance on paper but also what high-quality work looks like in practice.

How that training is delivered matters as well, particularly in distributed organizations with increased remote working options. Recording foundational training sessions allows writers across locations and time zones to learn consistently and at their own pace. Supplementing those sessions with live discussions, whether virtual or in person, gives writers an opportunity to work through real examples, ask questions, and apply concepts in context. This combination makes training more scalable without sacrificing the interaction, coaching, and mentorship that are critical to long-term development.

Organizations must also invest in the infrastructure that supports learning: repositories of redacted, high-quality example documents that set clear standards for new writers; templates and checklists that embed institutional expectations; consistent style guides that help maintain quality as teams grow; and robust mentoring programs. These investments are often less visible than technology initiatives, but they can more directly impact an organization’s ability to build and sustain writing capability over time.

Mentorship remains the hardest element to replace. Institutional knowledge does not reside in a standard operating procedure; it resides in the people who have spent years doing the work. Pairing newer writers with senior writers and creating structured opportunities to shadow experienced colleagues on active projects can accelerate development while preserving critical knowledge that might otherwise be lost through attrition.

Development programs must also evolve to address AI directly, focusing not only on how to use these tools, but also on how to critically evaluate their output. That includes recognizing unsupported or fabricated content, verifying information against source materials, and knowing when a passage requires deeper scrutiny than a tool can provide. Treating these capabilities as dedicated training competencies, rather than skills writers are expected to develop independently, will distinguish organizations that use AI responsibly from those that don’t.

This goes beyond HR’s remit, it determines whether an organization can build, sustain, and scale the expertise required to deliver on the flexibility, continuity, and quality that a hybrid model promises.

From filling a gap to building a bench

FSP relationships are most effective when designed to do more than fill an immediate resource gap. A study needs a writer, so a writer is placed, and in some cases that is where the relationship ends. The more effective approach goes beyond placement alone, incorporating structured training, mentorship, and enough continuity for expertise to deepen over time and strengthen the bench. The result is not simply a staffed project, but a stronger and more sustainable talent bench.

The tools will continue to improve, and AI will undoubtedly play a larger role in how medical writing work gets done. But technology alone will not solve the industry’s capacity challenge. The organizations that come out ahead will not necessarily be the ones with the most advanced AI capabilities. They will be the ones that continue investing in developing, mentoring, and cultivating the expertise that enable their teams to use those tools effectively. Ultimately, there’s no shortcut around talent, and building capacity means creating more of it.

Janice Worley, global head, FSP medical writing; and Lori Eads, senior director, FSP medical writing; both with ProPharma

References

1. Industry whitepaper (2024), as reported in IntuitionLabs, “FSP vs. Full-Service: A Guide to CRO Outsourcing Models” (2026).
2. The Regulatory Review, “The Governance Gap in Clinical AI” (August 2026), citing FDA and American Medical Association data on AI use and hallucination rates in clinical decision support.