Most AI Centres of Excellence fail for organisational reasons, not technical ones. The order in which you hire — and where the function reports — matters more than the algorithms you choose.
In life sciences, an AI CoE has to satisfy two masters at once: scientific rigour and production reliability. That dual demand shapes who you bring in first. Hire purely for research brilliance and nothing ships; hire purely for engineering and the science is shallow. The sequence below is the one we most often see work.
Why sequence matters
A CoE that hires data scientists before it has data infrastructure spends its first year cleaning spreadsheets. One that hires a research lead before a delivery owner produces impressive prototypes that never reach a clinic or a customer. The early hires should establish the foundations — leadership, data, and a route to production — before you scale the specialist bench.
“Hire the person who can ship a model into a regulated workflow before you hire your tenth person who can build one.”
The first five hires
1. CoE Lead / Head of AI: a credible leader who bridges science, engineering and the business, and can win executive trust. 2. Data / Platform Engineer — because models are only as good as the pipelines feeding them. 3. MLOps Engineer — to make deployment, monitoring and validation repeatable and auditable. 4. Domain / Clinical Lead — to keep the work anchored in real scientific and regulatory reality. 5. Senior ML Scientist — the deep specialist, hired once there is a foundation worth building on.
Where it should report, and the pitfalls
A CoE buried three levels down in IT rarely changes how the business operates. The most effective ones report close to the executive team with a clear mandate and a small number of high-value use cases. The common pitfalls are predictable: hiring specialists before foundations, chasing novelty over deployment, and isolating the team from the clinicians and commercial users it is meant to serve.
Where AI is being applied across the industry
The CoE you build should be pointed at where AI is genuinely moving the needle. Across life sciences, five application areas are pulling away from the pack — and a cohort of companies is defining each.
Drug discovery & design. AI is compressing the earliest, most expensive stages of R&D — target identification, molecule generation and protein structure. Recursion, Insilico Medicine, Isomorphic Labs (Alphabet) and Iambic are pushing AI-native discovery, building on the step-change set by DeepMind’s AlphaFold.
Clinical development & trials. AI is being applied to protocol design, patient matching, site selection and trial oversight. Companies such as Tempus, Saama, Unlearn (digital twins) and QuantHealth are reducing the time and cost of getting to readout.
Diagnostics & imaging. Machine learning now reads pathology and radiology at scale — PathAI, Paige, Tempus and Aidoc among those turning images and multi-omics into clinical decisions.
Regulatory, safety & medical affairs. Generative AI is automating document generation, pharmacovigilance and evidence synthesis. ArisGlobal, Causaly and Innoplexus are applying it to safety and scientific knowledge.
Infrastructure & foundation models. Underpinning all of the above, NVIDIA (BioNeMo), Ginkgo Bioworks and the cloud platforms are providing the compute and biology-specific models the rest of the industry builds on.
“The winners won’t be those with the most models; they’ll be those who embed AI into a real scientific or commercial workflow, and hire the leaders who can make that happen.”
For most organisations the opportunity is not to compete with these companies, but to partner with them, and to build the internal capability to deploy, validate and govern what they provide. That is precisely the leadership your CoE’s first hires must be equipped to do.
- Foundations first: leadership, data and MLOps before deep specialists.
- Anchor the team in domain reality and a route to production.
- Place the CoE near the executive team with a focused mandate.
Building an AI capability?
Talk to us →
Helping deliver talent that drives growth across the life sciences ecosystem.