Every growing biotech reaches the same decision about medical writing: build a team, hire an agency, or buy a platform and keep the work in-house. Each answer is right for someone. Each is expensive when chosen for the wrong reason.
Here is a plain comparison, written for the person who has to justify the choice.
In-house writers
The strongest option for continuity. Your own writers know the molecule, the history of decisions, and why the protocol says what it says. That knowledge is hard to transfer and easy to lose.
The weakness is capacity. Submission work comes in peaks, and a team sized for the peak sits underused in between. A team sized for the average falls behind exactly when it matters. Good writers are also hard to find and slow to hire.
An agency or CRO
The strongest option for flexible capacity. You buy the peak without carrying it all year, and a good agency brings experience across many programmes.
The weakness is context. Each new writer has to learn your programme, and the cost of that learning is paid on every project. Quality also varies by who you get, which is not always who you met during the sales process. Ask who will actually write your documents.
A platform with a smaller team
The option that has become realistic recently. Software drafts the descriptive, high-volume content and keeps it consistent with source, while a smaller team of experienced writers does the judgement and review.
The weakness is that it only works if the platform is trustworthy and the team actually uses it. A tool that writes fluent text without showing where each statement came from does not reduce review effort; it moves it. Source traceability is the property that makes this model work.
How to choose
Three questions usually settle it.
- How spiky is your workload? Very spiky favours an agency or a platform. Steady favours in-house.
- How much does programme knowledge matter? A single complex asset favours in-house. A portfolio of similar studies favours a platform.
- Where does the time really go? If it is mostly assembly and reconciliation, a platform helps most. If it is mostly expert judgement, it helps least.
Most organisations end up with a mix: a small core team, a platform for the volume, and an agency for the peaks.
Questions to ask any provider
Whichever route you take, a few questions separate good providers from expensive ones. Who exactly will write and review our documents, and what have they written before? How do you handle a change in source data halfway through a document? How do you keep terminology and numbers consistent across a submission written by several people? What does your quality check look like, and who does it?
For a platform, add one more: can a reviewer see where every statement came from without leaving the document? If the answer is no, review effort has not gone away. It has moved to the end, where it is most expensive.
What to avoid
Choosing on day rate alone. The cheapest writer is often the most expensive once rework and review time are counted.
And treating any option as a way to remove qualified judgement. Whichever model you pick, the conclusions, benefit-risk reasoning and causality still need someone accountable to sign them.
Our medical writing service and TraceDraft cover two of these models. If you want help working out which mix fits your pipeline, talk to us.
Where to go next
Explore GxP Copilot for AI-native validation, TraceDraft for source-traceable clinical documentation, or book a demo to see either on your own data.