Introducing the Pharma AI Solutions Library

After months of deliberation, brainstorming and cryptic LinkedIn posts, I’m finally ready to formally announce a project I’m calling the Pharma AI Solutions Library, whose goal is to help pharma decision makers identify the AI solutions and potential partners that most closely match their immediate needs and long-term strategy.

In this week’s post, I want to explain the reasons that I’m building it the way I am. But if you want to jump straight to what it is and how you can help me make it happen, check out the pre-launch sales page. Next week, it’ll be back to the regularly scheduled programming, speculating about AI/pharma partnerships and solutions.

Just over a year ago, I wrote a similar post to this one announcing a project called the Biotech Reference Stack that was meant to help biotech startup teams find the right data management and analysis software. That turned out not to have the impact I was hoping it would. But it did teach me a number of important lessons that shaped this new approach. So in today’s post I’m going to explain the four most important of those lessons, and what I’m doing differently this time as a result.

Lesson 1: Keep it simple

I wanted the Reference Stack to capture all the subtle details that are lost when you lump tools into convenient categories, so I created a framework that tried to transcend categories. But all it did was make things too complicated. Way, way too complicated. It was too much work for solution providers to figure out how their tools fit into the framework, and it was equally hard for decision makers to figure out how the framework represented their needs. In the end, it turned out categories are actually a good idea, even if they’re a little fuzzy.

The way I eventually came to terms with this was realizing that LLMs/AI can potentially solve this problem. Instead of trying to create a rigid framework that captures all those subtle details, I just need to capture all the subtle details, then let the LLM sort it out. (I’m glossing over some details, obviously…) So I can ask vendors to provide the same information in free-form, even stream of consciousness, and get similarly useful results, possibly even better, in a form that’s easier for decision makers to understand.

Lesson 2: Vendors are experts on their competitors

In the many conversations I had with solution providers, it became clear that they had all studied the landscape of solutions and knew the pros and cons of their competitors in much deeper detail than I ever could. I mean, it makes sense, right? That’s kind of their job.

But what was perhaps less intuitive was that if I asked them in the right way, in the right context, most solution providers (at least the engineers, if not the sales people) would objectively tell me where their competitors were better or worse. The hard part was creating that right context, but I’ve learned a trick or two (that I won’t give away here.)

Lesson 3: Sell the cake, sneak in the vegetables

The Reference Stack was focused on backend data management and analysis because that’s an important and overlooked aspect of data-driven biotech that often comes back to bite teams later. Just like vegetables. But if you want people to eat their vegetables, you first need to get their attention with something they want today. That’s the cake. And the cake is AI.

The Solutions Library is focused on AI because that’s both what folks are excited about and where they need the most help and guidance. In the course of getting those AI solutions to work, they’ll recognize that they need to clean up their data, etc. But they need to understand the AI solutions first.

Lesson 4: Make the chicken and the egg in parallel

With any two-sided resource like this, it’s a chicken and egg problem: To get solution providers to participate, they need to know that decision makers will use the resource. But for decision makers to use it, they need to know that all the main solution providers will be covered.

With the Reference Stack, I focused on the vendor side of the equation - trying to get them to provide information based on the promise that the users would come. But while most of the companies were interested in participating, they were also very busy. And yes, as I noted above, I made joining the Reference Stack unnecessarily difficult. But I also mostly ignored the decision-maker side of the equation that would’ve motivated the vendors.

So for the AI Solutions Library, I’m going to try and build up the user base in parallel with the vendors. And that’s where you, dear reader, come in: The reason I’m starting with pre-sales, before I’ve even created the first resource, is because I need something to motivate solution providers and independent experts to find the time to contribute the information that will make the Solutions Library work.

Every subscription that I get today will bring me one step closer to overcoming the chicken-and-egg problem by proving that decision makers want these resources and will use them.

I’m asking you to put some faith in me, that I’ll be able to make it happen. I’m determined to make it work one way or another. I would love your help to make it happen a little bit sooner.

So if you haven’t already, check out the pre-launch sales page and consider signing up today.