Picture two launch plans on the same desk in late 2025. Both have the same slide near the front: a revenue line that falls off a cliff as a big product loses exclusivity, and a stack of new launches meant to fill the hole. One plan assumes the field will absorb each new story the way it always has. The other assumes the field will have less time with each HCP, more launches to learn at once, and new tools in their pocket that nobody has quite figured out how to govern.
We think the second plan is the realistic one. As commercial teams build their 2026 plans, three shifts deserve more attention than they usually get: more launches with less room for error, fewer and more demanding HCP interactions, and generative AI moving from pilot to field. Each one changes what "a ready field force" means. This is our outlook as of October 2025, written for 2026 planning; it is a perspective, not a forecast, and the weight of each shift will differ by company, portfolio and market. Here is how we see them, and what we would do about each.
Shift one: more launches, and far less room for a slow start
The arithmetic is not subtle. IQVIA's Launch Excellence IX report, published in June 2025, estimates that about $200 billion of industry revenue is at risk from loss of exclusivity by 2030. To replace it, the industry will need an unprecedented number of launches, and they will need to perform consistently well, not occasionally well. How much of that lands on any one company depends on its exposure: some face a steep exclusivity cliff, others very little.
That second part is the hard bit. McKinsey's analysis of 210 launches between 2003 and 2009, published in 2014, found that about two thirds of new drugs missed pre-launch consensus sales expectations in their first year. Worse, the first year tends to set the trajectory. Of the drugs that underperformed in year one, 78% still underperformed in year two, and most first-year underperformers remained below expectations in subsequent years. Outperformers were stickier on the upside too, though less so: 65% stayed ahead in year two.
So a slow first year is not something a brand team can reliably fix later. And for companies with a crowded launch calendar, more launches mean more simultaneous first years, each competing for the same field force, the same launch budget and the same slots in the cycle meeting agenda.
There is a practical consequence that rarely makes it onto the strategy slide. When three brands launch in the same year, the field hears three kickoffs, three sets of key messages and three objection handlers. Each brand team, reasonably, thinks its story deserves the most time. The rep, reasonably, remembers whichever one was practised most recently.
What does that do to a field force? In practice it means a representative who used to carry one or two established brands and learn a new one every few years may now be asked to master a new indication, a new competitor and a label update within the same six months. In specialty and oncology, where approvals often come through accelerated pathways and indications expand quickly, the story may change again before the field has fully absorbed the first version.
What to do about it
Treat field readiness as a launch metric with the same standing as access, formulary wins and share of voice. That means defining, before launch, what the field must be able to explain in the first eight weeks, measuring it by person and by territory rather than by course completion, and reviewing it in the same launch meeting where you review early sales signals.
It also means accepting that a single launch meeting cannot carry the load. Knowledge from a kickoff fades fast. If the first year sets the trajectory, the weeks straight after the kickoff are the ones that matter most, and they need a deliberate reinforcement plan.
Shift two: fewer HCP interactions, and each one asks more of the rep
The rise of the informed, self-diagnosing patient has been discussed for years. That trend has not gone away, but its effect on commercial teams is indirect. What matters for the field is what it does to the HCP.
Many specialists now arrive at a meeting with a rep having already read the congress abstract, seen the competitor's headline and, quite often, fielded a question from a patient who read something online. In some of the markets we work in, access is tighter than it was, with more practices limiting rep visits and more of the relationship moving to remote, digital or MSL-led contact. This varies a lot by country, specialty and account type, and we don't have a single number for it. Our impression is that, where access has tightened, HCPs' expectations of each interaction have gone up.
When you get fewer minutes with each HCP, the cost of an unprepared rep goes up, not down.
This is where omnichannel thinking often misses a point. Organisations have invested heavily in channel orchestration: the right email, the right webinar invitation, the right sequence of touchpoints. That work is valuable. But the moments that move a prescriber are usually still conversations, whether in person, on a video call or with an MSL. The channel strategy decides when those conversations happen. It does not decide whether the person in them can answer the question that comes up in minute three.
A scenario we see often: a rep has a well-timed meeting with a respiratory specialist, set up by an excellent sequence of digital touches. The specialist asks how the product's exacerbation data compares with the biologic she already uses in a specific phenotype. The rep knows the headline, but not the subgroup, and not what the label allows them to say about it. The meeting ends politely. The orchestration worked. The conversation did not.
What to do about it
- Identify the ten or fifteen questions that decide whether a conversation moves forward in your therapy area, and make sure every rep can answer them from memory, within label.
- Test knowledge in the form it will be used: realistic objections and clinical scenarios, not just definitions.
- Close the loop with Medical. The questions that the field cannot answer are a signal about the questions HCPs are asking, and that is useful to MSLs and to the medical information team.
- Align the facts across Marketing, Medical and Market Access, each in its own role, so the HCP never hears three slightly different versions.
That last point is an old one, and it holds up. The organisations we see doing this well have a single source of approved content and a shared view of what the field knows, so a gap spotted by Medical can be addressed by Sales training within days, not at the next cycle meeting.
Shift three: generative AI moves from pilot to the field
Over the past two years, many pharma companies have run generative AI pilots. In our experience most were internal and low risk: summarising documents, drafting emails, searching policies. In 2026, we expect the pressure to be to put these tools in front of the field, answering product and disease questions during the working day.
It is easy to see why. A rep preparing for a meeting at seven in the morning cannot call medical information. A knowledge assistant that answers from approved material in seconds is an obvious help. And the technology has matured quickly.
The risks are just as obvious, and specific to our industry. A general-purpose model used without grounding in approved content can improvise an answer about an off-label use, or blend two trials into one convenient sentence. Used that way, it is not a productivity tool. It is a compliance incident waiting to happen. The regulatory framework is also moving. As of October 2025, the EU AI Act had entered into force in August 2024; its AI literacy obligations for organisations deploying AI have applied since February 2025, and obligations for general-purpose AI models since August 2025.
We think three design choices separate AI that helps the field from AI that creates new problems:
- It answers only from approved content: the label, the approved visual aids, the medical information responses. When the documents do not cover a question, it says so rather than guessing, and it points to the right channel, often an MSL.
- Experts sign off on what goes in, and conflicts between documents are flagged instead of quietly resolved by the model.
- It feeds back. The questions reps ask a knowledge assistant are some of the best signals you will get about what the field does not know and what HCPs are raising. That data should reach Training, Marketing and Medical.
There is a further point that gets less attention. An assistant that answers questions is not the same as a rep who knows the answers. During a call, the assistant can confirm a detail, but the rep has to own why the specialist should consider switching. AI in the field should reduce the time spent searching, not replace the need to understand. The best use we see is the two together: AI that answers at the moment of need and makes it easy to check, and a readiness program that makes sure the core story is held in memory.
A fair objection: "these are not new trends"
Fair. Launch pressure, access constraints and digital tools have been on every strategy slide for years. What is new is the combination, and the lack of slack.
A field force can handle one of these shifts with the tools it has. A heavy launch year can be managed with a big kickoff and extra training days. Tighter access can be managed with better targeting. A new AI tool can be piloted in one country. All three at once, with a revenue gap that does not wait, is different. It exposes the weakest part of most commercial models: the assumption that information delivered is information learned.
The companies that come through 2026 in good shape will, we suspect, be the ones that stop treating field knowledge as an input (hours of training, modules completed) and start treating it as an outcome they measure, person by person and topic by topic.
Where learning fits in all three
All three shifts land on the same place: the knowledge a person carries into a conversation. That is where we work. Atrivity runs short, competitive challenges between colleagues, battles in a synchronised window around kickoffs and cycle meetings, and Gap Fixing that sends each person the fix for their own mistakes. Readiness is reported by team, country and topic, and persistent misconceptions are flagged so managers can coach them. Atrivity Assistant answers only from approved documents, and the questions it receives feed back into the readiness program.
None of that is a strategy by itself. But if your 2026 plan depends on more launches going well, in fewer and harder conversations, with new tools in the mix, the question worth taking to your next planning meeting is this: how will we know, in week four, that the field is ready, rather than finding out in the quarterly numbers?