It was the fall of 2017, and I had just gotten a new gig that I was really excited about.

It was a chief of staff role within the Business Analytics group at Pfizer, a team of over 150 colleagues all over the globe.

Beyond the role itself, what I was really excited about was the innovation we had potential to bring to the whole company around how to leverage the two new, hot trends of the time: Big Data, and AI/Machine Learning.

I’d also just moved back to NYC and was energized by the digital health ecosystem taking shape around the city.

This was a time when Flatiron Health was growing its office in SoHo, Rock Health just established an east coast presence in Greenwich Village, and new digital health accelerators were popping up throughout the city.

In my spare time, I started meeting with startups, investors and others excited about this space.

The conversations were filled with all the benefits we can provide to patients and how the aggregated data - from EHRs, genomics, real-world evidence - could transform how we delivered care, and even how we discovered new therapies.

I couldn’t help but see the opportunity, especially for the young startups that I’d meet at accelerators, to partner with companies like Pfizer - and pharma more broadly.

I began trying to help those startups make inroads into the pharma industry by making introductions and helping them shape their offering and value proposition.

So, my inbox began to fill with pitch decks. And outside of work hours, I tried everything I could to help those companies kickstart their partnerships.

To be honest, it wasn’t easy. In fact, in most cases, it was an incredibly sobering experience, and something that stays with me to this day.

In most cases, I’d watch sales cycles stretch to 12, 18, 24 months… much longer than startup runways could sustain. I’d watch startups with impressive ML models or rich datasets search for problems that needed to be solved, rather than working backwards from “hair on fire” problems for their customers.

A lot of my time became education: helping startups understand how pharma actually worked and what kinds of insights would move the needle for us.

What I realized recently, is that while AI in 2026 does look different than AI a decade ago, many of the questions that stumbled ambitious startups of the previous era, remain the same:

While the technology has evolved, questions like this remain fundamental.

Having clear answers will not only help you find and grow your customer base, it will make it a lot easier to attract funding.

But before you can answer them, you need to know what kind of AI company you are in the first place.

Because ultimately, that will be the lens through which both pharma partners, and investors, will be evaluating you.

Let's start there.

Not all AI companies are evaluated the same way.

An AI-native drug discovery company faces completely different investor questions than a digital health startup using AI for care delivery.

A platform selling to pharma R&D teams gets scrutinized differently than a diagnostics company seeking FDA clearance.

The problem is, many founders haven't clearly defined which category they're in - or they're straddling multiple categories without realizing it. This creates confusion in investor conversations and makes it harder to tell a tight story.

Here's a simple framework to help you identify where you sit.

Start with one question: What is your primary output?

From there, the tree branches based on whether AI is core or complementary to your product, who your customer is, and what kind of biology you're working with.

This leads to seven distinct archetypes - each with its own investor dynamics, key questions, and positioning considerations.

Let's walk through each one.

Protein design, cell engineering, pathway optimization

What this means: You're using AI to engineer biology at a fundamental level - designing novel proteins, optimizing metabolic pathways, or programming cells to perform specific functions. Here, AI is enabling designs that wouldn't be possible through traditional methods.

Investor temperature: This category saw significant heat through 2023-2024, with landmark moments like AlphaFold shifting expectations for what's possible. However, investors have grown more discerning. The initial wave of “we use AI for protein design” pitches has given way to harder questions about validation, defensibility, and time-to-clinic. Investors are still excited, but they want to see wet lab validation, not just computational predictions.

Key potential questions from investors:

Your positioning move: Lead with wet lab validation, not computational benchmarks. Investors have seen too many pretty in silico results. Show them data from actual experiments, and be clear on whether you're an asset company or a platform.

AI-generated candidates, predictive models, novel target identification

What this means: Your company exists because of AI - the drug candidates themselves are generated or identified through machine learning models, not just accelerated by them. You're not a traditional biotech that added AI; you're an AI company that happens to make drugs.

Investor temperature: This category has cooled significantly from its 2021-2022 peak. High-profile clinical setbacks and the reality that no AI-discovered drug has yet reached market approval have made investors more cautious. The bar has shifted from "we use AI" to "show me the clinical data." That said, companies with differentiated approaches and real pipeline progress are still getting funded, just at more disciplined valuations.

Key potential questions from investors:

Your positioning move: Lead with clinical or late-preclinical assets, not computational benchmarks. If you're pre-clinic, be very clear on the specific bottleneck your AI solves and show evidence it's working in your own pipeline, not just on public datasets.

Biology-first companies where AI accelerates but doesn't define the science

What this means: You have a compelling biological insight or therapeutic approach, and you're using AI/ML tools to move faster - optimizing leads, analyzing data, identifying biomarkers. But if AI disappeared tomorrow, you'd still have a company. The biology is the moat; AI is the accelerant.

Investor temperature: This is actually a sweet spot right now. Investors are comfortable with the traditional biotech model and appreciate companies that use AI pragmatically without over-promising. You avoid the "AI hype" discount while still signaling operational sophistication. The key is not overselling the AI component while also not underselling legitimate computational advantages.

Key questions you'll get:

Your positioning move: Lead with the biology and clinical strategy. Mention AI as a capability that accelerates your timeline or improves your hit rates, but don't make it the headline. Investors should leave the meeting thinking "great science with smart use of tools", not "another AI company."

Imaging/pathology AI, companion diagnostics, liquid biopsy, biomarker platforms

What this means: You're using AI to detect, diagnose, or stratify disease - whether through analyzing medical images, pathology slides, genomic data, or blood-based biomarkers. Your output is typically information (a diagnosis, a risk score, a treatment recommendation) rather than a therapeutic.

Investor temperature: Cautiously optimistic, with strong differentiation by sub-segment. Radiology AI has become crowded and commoditized, with reimbursement challenges. Pathology AI and companion diagnostics tied to specific therapeutics are more interesting. Liquid biopsy remains hot but capital-intensive. Investors want to see clear regulatory pathways (FDA clearance/approval) and, critically, reimbursement strategies that actually work.

Key questions you'll get:

Your positioning move: Lead with the clinical and economic value proposition, not the AI. “We reduce time-to-diagnosis by 40% and catch 15% more early-stage cancers” matters more than leading with the technical details of your model.

Virtual care platforms, remote monitoring, AI copilots for clinicians, mental/behavioral health

What this means: You're using AI to deliver, augment, or scale healthcare services - whether that's a virtual care platform, a clinical decision support tool, remote patient monitoring, or AI-assisted therapy. Your customer is typically a health system, payer, employer, or patient directly.

Investor temperature: Mixed, with significant segment variation. The post-COVID digital health correction hit this category hard. Investors are wary of "vitamin" products with high churn and unclear ROI. However, solutions with demonstrated outcomes, strong retention, and clear payer/employer willingness to pay are still fundable. Mental health and chronic care management remain active areas, but unit economics scrutiny is intense.

Key questions you'll get:

Your positioning move: Lead with outcomes and unit economics, not features. Show that you've solved distribution (who buys and why) and that your product is sticky. If you have clinically validated outcomes, put them front and center. Avoid the "platform" framing unless you genuinely have multiple validated use cases.

Trial design optimization, patient recruitment, real-world evidence, regulatory intelligence

What this means: You're applying AI to make clinical trials faster, cheaper, or more likely to succeed - whether through smarter trial design, AI-powered patient matching and recruitment, synthetic control arms, real-world evidence generation, or regulatory pathway optimization.

Investor temperature: Quietly building momentum. This category doesn't get the same headlines as drug discovery AI, but the value proposition is concrete and measurable. Pharma companies have clear pain points (trials are slow, expensive, and fail too often) and are actively buying solutions. The challenge is that this is often a services-heavy business with slower scaling dynamics than pure software.

Key questions you'll get:

Your positioning move: Lead with quantified impact on trial timelines or costs, backed by pharma customer references. Be honest about the services component if it exists - investors respect hybrid models when the unit economics work. Show a path to productization and scaling beyond founder-led sales.

Platforms, tools, lab automation, data infrastructure, MLOps for bio

What this means: You're building the picks-and-shovels layer - tools that other biotech and pharma companies use to do AI-driven R&D. This could be data infrastructure, lab automation and robotics, AI/ML platforms for drug discovery, or specialized foundation models for biology.

Investor temperature: High interest, but increasingly competitive. This category benefits from being agnostic to which drugs succeed - you win if the overall AI-bio ecosystem grows. Investors like the recurring revenue potential and platform dynamics. However, the space is getting crowded, and differentiation is harder than it looks. "We're AWS for biotech" is a crowded pitch.

Key questions you'll get:

Your positioning move: Lead with paying customers and concrete use cases, not technical capabilities. Show that you've found a wedge where you're genuinely better than horizontal tools. If you have data network effects or platform lock-in starting to form, highlight them - but be honest if you're still early.

Now that you know the archetypes, the real work begins: preparing for the questions that follow. These two resources will help you move from “I think we're positioned well” to “I know exactly how to answer that”

70+ questions investors actually ask - organized by archetype. Pressure-test your positioning before you walk into the room.

[Get the Question Bank]

Not sure which category fits your company? Use this prompt with Claude or ChatGPT to clarify your positioning and identify potential mixed signals before investors do.

[Get the Diagnostic Mega-Prompt]

Look, I get it. Building an AI-related biotech is hard enough without having to guess how investors will perceive you.

My hope is that this framework gives you a clearer lens… not just on how to position your company, but on what questions to prepare for and how to tell your story with confidence.

I believe that in 2026 (just like in 2017), the founders that will get the most traction will not necessarily be the ones with the most advanced models. It will be the founders that will be able to clearly articulate what kind of company they are building, for whom, and in a way that will allow investors to color between the lines in the easiest possible way.

So, here's my challenge to you: Before your next investor conversation, answer this question in one sentence. "We are a [archetype] company that [specific value proposition] for [specific customer]."

If you can't do it clearly, that's useful information. It means your positioning needs work.

If so, nothing to worry about - always better to discover that now than in the middle of a partner meeting.

After all, that’s why we’re all here.

Now you know what to do 🙂

See you next week!

- Vadim

PS: Are you building in the AI + bio space and wrestling with how to position your company? Hit reply and tell me what you're working on. I’d be glad to be a resource.

PPS: If you have someone on your team helping with fundraising or know another founder who could benefit from being in this community - I’d love to include them. They can join us here: [Join the Community]