What We Learned Bringing AI to the Lab: Insights from Our First 200 Calls
In the past 18 months, I’ve had over 200 conversations with key decision-makers in pharma, chemicals & biotech companies. Here are 5 lessons I wish I’d known during my first call:
When we started Reactwise, we knew chemical process R&D teams were under pressure to deliver optimized processes faster, run fewer experiments, all while avoiding hazardous/dirty/expensive chemicals.
What we didn’t fully realize in the beginning was how complex adopting new software can be in this space.
Here are some early lessons that really shaped how we think and build:
- The science isn’t the only obstacle.
Even though we had several successful papers published, we came in thinking the biggest question would be: “Does this algorithm work efficiently enough?”
But the real questions were:
- Will IT approve this?
- Can the software pass our InfoSec requirements?
- How does this play with existing hardware & software?
In pharma, a great product is just step one. A secure, compliant, and trusted one is what actually gets used and makes a difference.
- Scientists are open but cautious.
The desire to innovate is there. Chemists and process teams are excited about smarter ways to optimize experiments. But they're also rightly skeptical of black-box tools.
They want to understand what's happening behind the scenes, make sure the predictions/recommendations make sense chemically, and they want confidence it won’t disrupt validated processes/workflows.
- The real competition is the status quo.
Most teams aren't choosing between us and a competitor. They're choosing between doing something new vs. sticking with Excel, Design of Experiments, and pure expert-driven workflows.
Change takes time, successful pilot results and trust.
- Simplicity is key.
Chemists employed in process R&D don’t have the bandwidth to fundamentally learn Bayesian statistics, chemical featurization, and machine learning in detail - even though these are very powerful skills.
A clean user interface, explainable results & tooltips, and low-friction onboarding are what actually drive adoption.
- Trust builds slowly, but it matters most.
The first 200 calls taught me that trust is not a byproduct of sales - it is the sale.
It’s built through showing up consistently, understanding the realities of regulated environments, and being honest about what the product can and can’t do.
To everyone who gave us their time in those early conversations: thank you.
You didn’t just challenge our product - you made it better.
If you’re leading process development, digitalization, or innovation in pharma and want to reduce development time or explore how AI can support your chemists - we’d love to show you what we’ve built.
We’re already working with leading pharma and CDMO teams, and we’re always open to starting the conversation early.
Drop me a message to alexander@reactwise.com - let’s talk chemistry, optimization, and what’s next.
