TLDR: Commercial pharma digitised faster than its skill mix. Five capabilities now decide whether a team ships — orchestration, data literacy, martech ownership, AI fluency and compliance-by-design — and the constraint is how they are distributed, not how many people hold all five.
Omnichannel orchestration breaks at the data layer long before it breaks at the creative one
Commercial teams rarely struggle to produce content for multiple channels. They struggle to make those channels behave as one conversation, and the failure almost always originates somewhere unglamorous — in whether the organisation can tell that the physician who opened an email on Tuesday is the same physician a representative visited on Thursday. Orchestration is the capability that closes that gap, and it is a coordination and judgement skill supported by systems rather than a creative one. Naming it that way matters for hiring, because a team that files orchestration under marketing recruits for portfolio and storytelling, then discovers the appointed person cannot answer the only question the role exists to settle: what this customer should receive next, and when.
The dependency runs in a specific order. Sequencing content requires knowing what a contact has already received, which requires a single identity for that contact across field, digital and medical systems. Where identity is fragmented across three databases with different keys, no amount of campaign planning produces coherence, because each channel is optimising against its own partial history. Teams that invest in creative variety before resolving identity typically discover this eighteen months and a substantial budget into an omnichannel programme, which is an expensive way to learn a sequencing constraint. The order is not a preference but a dependency: every downstream capability consumes the identity layer, so work done above an unresolved one produces output that cannot be assembled into anything coherent later.
Orchestration also demands authority that most organisations forget to grant. The person accountable for the coordinated experience must be able to tell a brand team to delay a send because a field cycle is running, or to overrule a channel owner whose metric improves while the overall experience degrades. Without that authority the role becomes a scheduling function, and every channel continues to optimise locally. Naming the accountability in the org design costs nothing and determines whether the capability actually operates. The reason authority is the binding constraint is that the orchestrator’s job consists almost entirely of imposing costs on people who are being measured on something else, and goodwill runs out well before the second brand team is asked to wait.
Judgement about restraint completes the skill. The technically available frequency across owned channels usually exceeds what a specialist audience will tolerate, and healthcare professionals disengage quietly rather than complaining. Strong orchestrators plan the pauses as deliberately as the touchpoints, and they measure sustained engagement over a cycle instead of response to a send. That instinct is learned from running programmes and watching audiences fatigue, which makes it hard to teach in a workshop and worth paying for in a hire. The quiet nature of the disengagement is what makes experience irreplaceable here, since a team reading per-send metrics sees nothing wrong until the audience has already gone — and coordination of this kind rests entirely on being able to see what is happening.
Data literacy and martech ownership decide whether the stack pays for itself
Most commercial pharma teams already own capable technology and capable reporting. The gap sits between the two: reports that describe activity without informing a decision, and platforms configured by someone who left, running workflows nobody currently understands. Both gaps close through people rather than through further purchasing. This is an uncomfortable finding for an organisation mid-way through a technology programme, because the remedy has no vendor attached to it and no implementation date, and it competes for budget against a platform upgrade that is far easier to describe in a steering committee.
Data literacy sets a lower bar than the job market implies. It means reading a performance report, forming a specific hypothesis about why a number moved, and testing that hypothesis without commissioning an analyst — a chain that requires curiosity and basic statistical caution rather than modelling expertise. The commercially decisive part is the hypothesis step, because a marketer who notices that engagement fell in one specialty and asks whether the content mix changed there will find causes the dashboard never surfaces. Teams without that habit generate reporting volume and decision paralysis simultaneously. Dashboards answer only the questions someone anticipated when building them, so the capability that matters is the one that produces the unanticipated question.
Martech ownership means someone inside the commercial team understands the stack at configuration level. That person knows how consent flags flow from capture to campaign selection, which fields the CRM treats as authoritative, and what an integration will silently drop. The reason this must sit in the business rather than purely in IT is that most compliance failures in digital campaigns originate as configuration decisions: an audience segment built on a stale consent field, a personalisation token pulling from an unapproved source. Only someone who understands both the workflow and the regulatory intent spots those before a send. An IT owner sees a technically valid configuration, and validity is exactly what a misconfigured audience has — the error lives in the intent, which is invisible from inside the system.
The two capabilities reinforce each other in a way that argues for concentrating them. Someone who configured the campaign knows exactly why an engagement number is unreliable, because they know which field it came from and what happens when a record is merged. Splitting the pair across two junior roles produces analysts who trust bad data and operators who cannot explain their own outputs. Where budget allows one senior hire rather than two, this pairing is usually the better concentration. Seeing clearly and building correctly still leaves the question of judgement — where a person has to intervene in work that machines and processes would otherwise complete unsupervised.
AI fluency and compliance-by-design are the same discipline viewed from two directions
These two capabilities appear on opposite ends of most skills frameworks, one filed under innovation and the other under risk. In a regulated commercial team they resolve into a single competence: knowing precisely where human judgement has to enter a workflow, and building the process so that it does. Filing them separately has a practical cost, since it puts them in different budgets, different training curricula and often different reporting lines — with the result that the people learning to use generative tooling and the people who understand where the claim boundaries fall are rarely the same people, or even in the same conversation.
AI fluency in this setting is mostly the skill of directing and checking. Generative tools produce fluent, confident, well-structured copy, and fluency correlates weakly with accuracy — which is precisely the failure mode a regulated environment cannot absorb, because a plausible claim that exceeds the approved label is worse than an obviously wrong one that review catches immediately. The practical skill is knowing which tasks tolerate automation (variant generation, summarising engagement data, first-draft adaptation) and which require verification against a source document every time. That distinction is a judgement about consequence, and it comes from understanding the regulatory context rather than the model. It also has to be made task by task, since the same tool is safe in one step of a workflow and unsafe in the next.
Compliance-by-design applies the same logic upstream. Content and data handling in this sector operate under national promotional codes, the oversight of Swissmedic and the EMA, the MDR and IVDR frameworks for devices, and the GDPR alongside the revised Swiss FADP for personal data. Teams that treat review as a gate at the end of production discover structural problems when changes are most expensive, since a claim issue found at asset stage means re-shooting rather than re-wording. Teams that build the constraints into the brief, the modular content model and the data architecture ship faster with fewer cycles — the compliance benefit and the speed benefit are the same benefit. That identity is the argument to make internally, because a control presented as protection gets deferred and a control presented as throughput gets funded.
The connection becomes concrete where AI enters regulated content production. A team that has already mapped which claims are approved, which sources are authoritative and which decisions require a named reviewer can introduce generative tooling into that map safely. A team without that mapping is adding volume to an unclear process, and volume is what turns an unclear process into an incident. Sequencing matters: the organisations using AI most productively in commercial pharma tended to have their content governance in order first. Five capabilities, then, each individually learnable — and the difficulty most teams actually face is arranging them across the people they have.
| Capability | Where it should sit | Symptom when it is missing |
|---|---|---|
| Omnichannel orchestration | One accountable owner with cross-channel authority | Channels perform well individually; the experience is disjointed |
| Data literacy | Distributed across the whole commercial team | Reporting volume grows; decisions do not change |
| Martech ownership | Inside the commercial team, not only in IT | Nobody can explain why a workflow behaves as it does |
| AI fluency | Content and campaign roles, with named review points | Fast drafts that fail review on substance |
| Compliance-by-design | Built into briefs and content models, owned by all | Late-stage rework and slipped launch dates |
Team capability is a distribution problem, which is why the unicorn requisition keeps failing
Reading the five capabilities as a person specification produces a job description no available candidate matches, and a search that runs for nine months before being rewritten. Reading them as a team-level inventory produces a solvable problem, because the capabilities have different natural homes: some belong to one accountable owner, others need to be spread thinly across everyone. The shift is inexpensive to make and rarely made, largely because a requisition is the artefact a commercial team knows how to raise, whereas a capability inventory has no template, no approval route and nobody whose job it currently is to own.
Start by auditing what the team currently holds against the symptoms rather than against the CVs. Slipped launch dates and late rework indicate a compliance-by-design gap. Growing dashboards with static decisions indicate a literacy gap. Channel metrics improving while overall engagement flattens indicates missing orchestration. This diagnostic ordering matters because the symptom identifies which capability to buy first, and buying in the wrong order wastes a hire — a data specialist added to a team whose real constraint is content governance will produce excellent analysis of a broken process. Symptoms are also the only evidence available that describes the team as it operates, since a CV describes what someone was hired to do rather than what they have since ended up doing.
Professionals should read the same inventory as a route map. Building from a strong existing base — commercial marketing, digital delivery, or a scientific or medical affairs background — and adding the adjacent capability that currently blocks your work compounds faster than collecting certifications. The people who end up holding several of these at once are what Edward Galle calls brand-tech specialists, and the strategic judgement those roles reward can be accelerated through HBR-backed training journeys rather than learned entirely on the job. Adjacency is the operative filter, because a capability added next to an existing strength gets used immediately and therefore consolidates, while one added at distance decays before an opportunity to apply it arrives.
The pressure behind all of this is durable. The World Economic Forum’s analysis of shifting core skills points to substantial turnover in what roles require across this decade, and regulated commercial functions sit inside that change rather than adjacent to it. Meanwhile Swiss life sciences hiring trends show persistent scarcity in specialist commercial roles, which means the supply side will keep rewarding employers who develop capability internally alongside hiring it. Those two pressures point the same way: a target profile that keeps moving makes adaptability the durable thing to select for, in a hire and in a development plan alike.
The teams pulling ahead treat this as an annual capability review rather than a recruitment exercise: audit the inventory, name the single binding constraint, then decide whether to develop it or buy it. Doing that once a year is a modest commitment against the cost of a stalled launch. The version worth running first is deliberately small — one meeting, the symptom list, and a single named constraint carried into the next planning cycle — because a team that produces one honest answer this quarter has something to test against next year, which is the point at which the exercise starts compounding.
Edward Galle sources commercial talent for pharma, MedTech and life sciences employers across Switzerland, the EU and the US. Employers can discuss an open role with our team ou see how our search process works; professionals building these capabilities can submit a CV for upcoming briefs.
References
- World Economic Forum. Future of Jobs Report. https://www.weforum.org/publications/the-future-of-jobs-report-2025/
- Panda International. Swiss Life Sciences Hiring Trends for 2026. https://www.panda-int.com/en-ch/insights/swiss-life-sciences-hiring-trends-for-2026/
