Most life sciences commercial teams have access to AI. Far fewer have changed how they work because of it. This panel included four people who have gone well past that line.
Ann Brine, Head of Marketing Services at Promega UK. Matt Wilkinson, founder of Strivenn, Ltd. and the synthetic customer evangelist. Nick Clare, co-founder of Succession, who spent 17 years at the lab bench before building one of life sciences' more interesting AI-native sales businesses. And Sarah Neely, VP of Strategy at Fizz and founder of SignalSwell, a consultancy built at the exact point where human word-of-mouth meets AI discoverability.
What followed was an hour discussion about what is actually happening inside life sciences commercial teams right now, where AI is making a genuine difference, and what it takes to get there.
Here are the lessons that the hour taught us.
Context is everything. Prompts are just the surface.
Nick Clare said it early and the panel kept coming back to it: AI without context gives surface-level results.
Nick's team at Succession has spent recent months making sure AI touches every single process in their business. Not as a one-off tool, but rather as a workflow layer. One of his recent projects involved building a skill inside Claude that lets a set of agents review any client they're working with, scan every published paper in that client's scientific area, and return the equivalent of a five-year postdoc knowledge base. Every commercial conversation Succession has now carries genuine scientific depth underneath it.
That is a fundamentally different proposition to asking Claude to write a follow-up email.
Matt Wilkinson drew the same thread through his synthetic customer work. The difference between a persona that tells you something useful and one that reflects your assumptions back with a job title attached comes down entirely to the quality of context you feed it. When he tested potential front covers for his upcoming book, his synthetic customers and a live LinkedIn poll came back with strikingly similar results: the three options weren't quite there. New options were built, and one won the synthetic customer panel. Book cover development, tested and refined.
The lesson is not that the tools are clever. It is that the tools are only as useful as what you put into them.
AI discoverability is not an SEO strategy. It is something new.
Sarah Neely works at a point most life sciences marketers haven't reached yet: the space between human conversation and what AI can find about your brand.
Her argument is straightforward. A potential buyer asks a colleague for a recommendation today. Tomorrow, they validate that recommendation with AI. If your brand's digital footprint doesn't reflect what your best customers actually say about you, you are not part of that second conversation.
Nick added a number that reframes the whole content question: the majority of sources cited by AI systems come from content published within the last 13 weeks. Content published within 30 days is three times more likely to be cited in an AI answer. It is no longer about building a library. It is about staying current enough to be visible in a window that closes fast.
Matt's team has already acted on this. They built a microsite for a client mid-website rebuild, loaded it with FAQ content structured around the specific questions their buyers ask, got the schema markup right, and within a week those answers were appearing in Google's AI-generated results. Answer the right questions, keep the content current, and the visibility follows.
"We have agency in this crazy robot world to tell our stories in the way that we want to tell them."
Brand is not a casualty of AI. It is the last defensible advantage.
The panel spent time on what Matt called AI slop: the generic, undifferentiated content that now floods every channel because everyone has access to the same generation tools. Ann's observation was pointed. A colleague had received content from a customer that could have been written for Promega. The customer's brand wasn't in it. Their differentiation wasn't in it. It could have come from anyone.
Matt's approach to mitigating this is worth adding to your "swipe file": take the Wikipedia page that catalogues all the known signs of AI-generated writing, pull out every telltale pattern, and build an automated check that runs your content against that list before it goes anywhere. Use the AI's own documented weaknesses against itself.
But the deeper point is one Ann made and Matt extended. Brand increasingly lives in your people, not your assets. The organizations that will hold their ground are the ones that enable their scientists, their field application specialists, their marketers and sales teams to tell real human stories from their actual experience. Matt referenced the concept of a trust portfolio, where organizations treat their people as distributed media channels rather than leaving all brand voice to a central marketing function. Ann made the same point from a different direction: staff as champions is the most cost-effective marketing a life sciences organization has. Word of mouth where it actually happens. Sarah's entire practice is built on making that kind of human conversation visible and durable in an AI-first world.
The gap between experimenting and committing is a management problem.
Ann described what sustained AI adoption took inside Promega: giving people time to learn, not just licenses to use. This could look like AI brought in as a welcomed partner rather than a performance pressure. Internal champions used as train-the-trainers rather than relying on a single central training event.
Nick flagged a specific failure mode his team had run into: getting so excited about a new tool that you throw it at your team before they've had time to get comfortable with the previous one. The result is a team permanently on a learning curve that never reaches the depth where the tool actually becomes useful. His principle: test it thoroughly, like a scientific experiment, before you embed it. Run it in parallel. Watch the results. Then make it part of the process.
Matt called on John Cleese's framework for creativity: there is a time for play, and a time for disciplined creation. The teams pulling ahead have made space for both.
The human in the loop is not a safeguard. It is the whole point.
Ann raised the question that stayed with the room: how do the next generation of marketers and sales professionals develop the editorial judgment that makes AI useful, if AI is doing the work that used to build that judgment?
Nick was the most optimistic. He pointed to 100 years of headlines warning that the next invention would end work, and noted that humans have always found new roles within new structures. Sarah's version of the same concern was quieter: as a mother and a business person, she called it a wild ride.
Matt drew the thread forward. The gap is not the tools. The gap is the students who learned in systems that said no to AI while the workforce itself was saying yes. Those are the people who need the most support, and they deserve it.
But across all four panelists, the operating principle was consistent: you are responsible for whatever goes out under your name. Nick's phrase was the one that stuck: treat AI like a genius intern idiot. You would not hand your intern the keys and walk away. You set the brief, you review the output, you make the call. The people in the loop are not there as a backup in case something goes wrong. They are the reason it goes right.
The recording of this session is available on the SAMPS website. If you have questions for the panel that didn't get answered, send them to info@samps.org and we will make sure they reach the right person.
