Norstella has announced the launch of Atlas, an agentic AI platform designed to support the teams tasked with making pharma’s most important decisions. It’s a critical component of Norstella’s approach to serving customers, so we caught up with Kris Kaneta, Chief Product & Innovation Officer at Norstella, to talk about what makes Atlas different, why the Clarity Test matters, and what the launch of Atlas CI signals about the future of AI in life sciences.
Kris KanetaAtlas is our agentic AI platform, and the simplest way to describe it is this: it turns 30 years of curated, structured, expert-validated data into finished work. Not search results. Not retrieved documents. Finished outputs: competitive landscapes, feasibility assessments, launch plans, access strategies.Those are decisions where a confident-sounding answer and a correct one are very different things. Atlas is built around that distinction.
What makes Atlas different from other AI platforms across life sciences?
Kris KanetaThe honest answer is the data foundation underneath it. Norstella has spent more than 30 years building a proprietary data network spanning clinical development, regulatory, commercial, and market access, linked at the entity level across the full drug lifecycle. That’s not something you can replicate by pointing a model at public sources.Most AI tools in this space are essentially wrappers. They retrieve information faster, which has value, but they’re still retrieval tools. Atlas agents reason across the entity linkages, workflow logic, and domain judgment built into Citeline, Evaluate, MMIT, and Panalgo over decades. That’s what allows Atlas to deliver decisions where a general-purpose model delivers descriptions.
You’ve described this in terms of a “Clarity Test.” What does that mean?
Kris KanetaThe Clarity Test is how we keep ourselves honest. It’s our internal benchmark for whether an output is genuinely fit to act on when the stakes are high. Every answer Atlas produces must meet four criteria:• Cited: traceable to a source. You can audit it.• Consistent: grounded in a stable context layer, not a rotating public corpus that shifts underneath you.• Contextual: matched to the actual decision workflow, not a generic response to a generic prompt.• Consequential: actionable, not just confident-sounding.General-purpose AI can clear one or two of those. Atlas is built to clear all four.
Atlas CI is the first agent to go live in Atlas. Tell us about it.
Kris KanetaAtlas CI is built for competitive intelligence teams. It compresses the time between a question and a finished, decision-ready output: competitive landscapes, drug profiles, catalyst timelines, executive briefings. The kind of outputs that today take analysts days or weeks to produce.What distinguishes it from other CI tools is that it’s built around the person doing the work and the workflow they actually follow, rather than around a data set. Each output carries its sources and is grounded in Norstella’s mastered data and the customer’s own internal intelligence. Atlas CI reasons across multiple discrete data sets that until now wasn’t possible. Most importantly, it doesn’t replace human judgment; it lets people apply that judgment far faster and more effectively.
What comes after Atlas CI?
Kris KanetaGreat question! We have dozens of agents operating across Norstella, and Atlas CI is just the first agent within Atlas and our team is busy working on many more, each of which is tailored to a specific role or persona in the industry. We have a ton of ideas about future agents, but the ones that users can expect to see in the coming months cover roles in business development and licensing, feasibility, asset and portfolio strategy, and protocol design. They’re all built on the same platform, grounded in the same data, and they’re held to the same Clarity Test.The key thing is this: our agents are shaped around the person doing the work and the workflow they follow. We’re building for pharma’s highest-stakes decisions, and that’s a different brief from building for general productivity.
How has Norstella been using AI until now?
Kris KanetaAtlas is a huge project, and one we’re really excited about, but we’ve been leading the AI space for years. We have dozens of what we call “orchestrating agents” which not only power Atlas but also accelerate our own product development and support the ongoing curation of the data itself. These include Evaluate Omnium, Norstella LinQ, and the award-winning Ella solution.Part of the reason we’ve been in the AI space for so long is the sheer volume and complexity of our proprietary data. We are constantly refining millions of life sciences data points across the drug development lifecycle: 10 trillion transactional data points, 10 billion real-world data points, 38 billion clinical data points, 25,000 primary research engagements every year, and over 3,000 brands across 400 indications. Great AI is the only way to tame this resource and deliver real value from it.
Norstella Atlas combines Norstella’s proprietary data assets with purpose-built AI agents to deliver decision-ready outputs across the drug development lifecycle.
Want to learn more? Explore Atlas today →