Read time: 5 minutes

Welcome to the Succession newsletter where 2,000+ life science sales reps improve their skills in 5 minutes per week. If you’re getting value from these newsletters, we'd love it if you could forward it along to your sales colleagues. If you’re new here, subscribe below.

Succession Bio works with life science/biotech companies to help drive sales, licensing, and partnership opportunities.

We do this through market research to identify the right companies and people, craft scientifically credible messages, and then perform the outbound sales and marketing tactics on your behalf to facilitate meetings with the right people at the right companies.

Succession

• Specializes in life sciences/biotech (it's all we do!)

• Provides market research, messaging, and outbound sales/marketing services

• Facilitates meetings and opportunities with the right people at the right companies for our clients

• Sales training for teams of 10+ who want to find and close more deals with biotech and pharma

A few points up top: we’ll be at ELRIG Drug Discovery in London next week, so if you’re around, come see us at booth F07 for some good non-conference coffee, message us if you want to grab a drink, or grab some time to meet here.

If you’re around on set-up day, join us at the Hype-to-hands-on AI workshop (details above, registration here) for some practical wisdom, networking and mini workshops led by life science sales and marketing experts.

Ok, let’s get to it.

Going AI-first as a life science commercial team

On Wednesday, we ran the first session of our six-part webinar series on putting AI to work in a life science commercial team (series here in case you want to check it out). Harrison had gone to France. His Wi-Fi had not. Nick spent the first 30 seconds talking to us on mute. Nobody could share their screen. "We're a technologically advanced company," Nick told a live audience, and then his video cut out.

Once we were actually running it was a genuinely good hour, so I thought I’d give you the summary in case you missed it.

You need two things to make this work

The premise is simple. Most teams use AI the way you'd use a very clever intern who lives in a browser tab. You ask it something, it answers, you go about your day. Going AI-first means moving from that to something the whole team runs on. Nick covered the people half. Harrison covered the building half. Internally, Nick is known as the human one, and Harrison is the one who talks to the agents, so the split was not exactly a stretch, but it’s a good reminder that you have to think about both sides.

Part one: the people

Nick's opening line was the best summary of the whole session: "AI-first is not something you inflict on people."

It has to be built from both ends. Leaders hand down three things: the tools and the time to try them, permission to get it wrong, and a say in how the work gets done. The team hands back three things: better ways to do their own work, an honest account of what worked and what didn't, and the changes worth rolling out to everyone. Miss either end and it doesn’t work. You get expensive tools nobody opens, or brilliant ideas that live and die in one person's head.

Give people room to try

This is about fear. If people get judged for trying something that didn't work, the experimenting stops, and experimenting is where the magic happens. So a leader's job is to remove that fear on purpose. Say out loud that some attempts will go nowhere, and that’s ok, the point is to experiment.

One thing here that gets underestimated is budgets. If you have a budget for anything, have a budget for tinkering. Everyone on the team should feel encouraged to sign up for a new tool, burn some credits, and see what happens.

Make what they find count

Someone finds a better way, a prompt, a tool, a shortcut, a new process. They show the team. Is this useful to the team’s way of working or does it make them better, or faster? If yes, it gets integrated into how everyone works.

Having an approval step here is important. Have a process for new stuff. Test whatever got built, work out how it fits with everything else, and decide whether it's safe to roll out without breaking everything. 

Two filters, then: is this just for me or for the team, and will it break something.

Let each person work their own way

Everybody’s brains work differently, so the way they work with AI systems will need to be different to get the most out of it. Whether that’s how to visualise outputs, how to kick off new workflows, manage tasks or execute creative ideas.

The lesson here is to give people the freedom to use the systems you’ve built for the team, in a way that works for their brain. And where they need to, build new things that help them interact with AI in a more productive and enjoyable way.

One recent example Nick gave in the session involved what we call at Succession a “tactical squirrel”. Essentially getting distracted and going off on a side quest with AI, except it actually produces something useful (highly encourage everyone to do this).  

We have a lot of agents working in shared Slack channels. The output was great, but I was losing my mind going in and out of a dozen threads to see what each one was doing. It worked fine for Harrison’s brain. It did not work for me. So I built an interactive 3D world where every agent sits as a little robot and you can see at a glance what each one is working on, which one’s need your input, and which client it's for.

So now we have two ways to visualise and do the same work. Nobody was forced to use either.

Part two: the building

Harrison opened with "now that we're done with all the fluffy stuff," which tells you everything you need to know. 

His section was six pillars, with one goal sitting on top. Here they are in order.

1. Context

If a new hire can't learn who you sell to and why you win from one place, neither can your AI. Context is the shared record of what you sell, who buys it, what they care about, your personas, your case studies, your objections and your competitors. Build it once, and every AI task starts from it instead of from a blank page.

Practically, this can be a proper database if you have someone technical, a project in whatever AI tool your company pays for, a shared Drive folder, or a GitHub repo if you want to push the boat out. Take everything you already have lying around in slide decks and brain-dump it in.

The point that unblocks most people: you don't need the crown jewels to start. Most sales teams can be effective on public-facing information alone. If you work somewhere large and the CRM is locked down, start with the decks.

2. Data and access

Your AI can only use what it can see. If it can't see your CRM, your call recordings, your email and your calendar, it will never be as useful as it could be. Harrison was blunt that this is the biggest sticking point for most teams right now, and that it's a leadership job to fix it.

The reason it stays stuck is that people ask IT for the data without explaining the outcome. IT sees risk. Nobody shows them the benefit. Make the business case: a rep prepped for a call without 30 minutes of clicking through Salesforce is the benefit. Harrison's line about "the 17 different button clicks you have to go through to find the thing you're looking for, if it's even in there" got a knowing silence from the audience.

He also made a distinction I liked. Everyone talks about AI as efficiency, doing the same thing faster. The real prize is effectiveness: a rep who walks into the call knowing more, with better information, is simply a better seller.

3. Shared playbooks

A playbook (or a skill, if you’re familiar) is written instructions for one task. How we do account research. How we write prospecting messaging without the tells that scream AI. How we build a territory plan.

Without shared ones, two reps asking the same question get two different answers. Ten reps, ten prompts, ten results. With them, a new rep on day one could run the territory plan playbook, then the account research playbook on every account in it, and have something usable before lunch.

The easiest way to build one: do the task manually once. Ask for account research on Pfizer's oncology programmes. Refine the output until it's what you want. Then say "now write me a skill file so I get this every time." 

The models are good enough now that you're better off describing what a good output looks like than listing every step of the process.

Nick jumped in here with a nuance worth repeating. His whole section said "everyone works differently." Harrison's said "everyone follows the same process." Both are true. The process and the data stay the same. How you consume the output is yours. Don’t like visual or written outputs? Get AI to turn it into a podcast instead. Same answer, different door.

4. Single player versus multiplayer

Single player is you and a chat window. It's great for call prep, drafts and research, and it stops the moment you close the laptop.

Multiplayer is agents living in Slack or Teams, where several people and several agents can work in the same thread. One team member kicks off some work, then tags another to review it when it's done, or picks it up from wherever it got to. With teams that sit across time zones, this is the difference between work that waits and work that keeps going.

5. Guardrails

Three levels. Level one runs on its own: account research, enriching contacts, updating a CRM field, putting a draft in your drafts folder. Level two gets reviewed by a person: lists, email drafts, account plans. Level three needs a named person's approval: anything a buyer or client sees, pricing, scientific claims, launches.

It doesn't have to be all or nothing. Give the agent read access to the whole CRM and require sign-off on any write. One data rule to hold without exception: client data under NDA never goes into a personal AI account.

6. The learning loop

This is how the system gets better without anyone babysitting it.

Launch a campaign to three personas with three messaging angles. The agent watches the replies. Variant B worked best for persona C. That finding goes back into the playbook, so the next campaign aimed at persona C starts from variant B without anyone opening a reporting dashboard. Or: you handle an objection well on a call. The transcript gets analysed, the handling goes into context, and the next person's call brief suggests it.

Then the compounding argument. Twenty reps, every learning feeding back. A team that builds this today versus a team that starts in a year: the first one has a year of improvements baked in before the second one has even started.

The roof: selling time

The goal on top of all six pillars is more active selling time on qualified opportunities. Salesforce's State of Sales 2026 stat puts rep selling time at around 40% today.

Every sales tool in history has promised to give reps time back, and reps still drown in admin. The difference now is that the admin genuinely can go: call prep, research, account plans, QBR decks, the technical questions that land on a rep who should be selling. But removing admin alone doesn't raise selling time. If a rep has five calls a week and no admin, they have five calls a week. You also have to feed the top of the funnel so they have more qualified opportunities to spend that capacity on. Do both and a rep goes from working ten qualified opportunities at a time to fifteen or twenty.

How to measure it: pull every rep's calendar and count the hours actually spent with customers and prospects. Move that number up. Pair it with qualified pipeline created so nobody games it by padding their diary.

Score yourself

Seven yes-or-no questions, one per pillar. Your weakest answer is where to start.

1. Leadership: is one named person responsible, with a goal you review every month?

2. Context: could a new hire learn who you sell to and why you win from one place?

3. Data and access: can your AI see your CRM and calls, and build and enrich a list itself?

4. Shared playbooks: would two reps asking for the same task get the same quality back?

5. Multiplayer: does your team share AI it works on together, beyond personal chats?

6. Guardrails: is it written down what the AI can do alone and what needs sign-off?

7. Learning loop: when something works, does everyone's AI improve without manual updates?

And finally

If you didn’t join us on Wednesday, you can catch the recording on YouTube (link below). For those who did register, the recording is also at the same link you used to join, and it'll be there whenever you want it. 

Session two is on October 29: "Your best rep's brain, written down," which is the deep dive on context and playbooks. By the end of the series, you’ll have a whole sales pipeline's worth of AI builds you can take back to your team, so if that’s something you’re exploring right now, register for the series here.

Until next time,

Jess

Webinar recording:

1. Lead Generation: We’ll build target lists, write scientifically relevant messaging, and send messages on your behalf to book qualified sales meetings with biotech and pharma companies.

2. Training for Teams: If you want to upskill your team around prospecting, driving to close, key account management, AI, or any other topic, we can put together a training plan specific to your organization’s needs.

3. Strategy Call: Need more than training? Want help implementing and executing your sales strategy? In a 30-minute call, we will assess your company’s current situation and identify growth opportunities.