Norstella has launched an agentic AI platform designed to combine proprietary life sciences data with purpose-built AI agents to produce decision-ready work across drug development.

The platform, Norstella Atlas, is intended to support tasks including competitive intelligence, feasibility assessments, launch planning and market access strategy.

The company said Atlas builds on proprietary clinical, commercial, market access and real-world data from its Citeline, Evaluate, MMIT and Panalgo brands.

Rather than positioning AI as a tool that simply answers questions, Norstella is developing Atlas around specific business roles and workflows. The first persona-specific agent, Atlas CI, is now available to support competitive intelligence teams working in pharmaceutical and biotechnology companies.

The launch reflects growing interest across the life sciences industry in agentic AI systems that can carry out multi-step tasks and generate usable outputs, rather than simply producing individual responses to prompts.

Norstella targets decision-ready AI outputs

Norstella said Atlas is designed to turn its connected data assets into finished work, with the initial Atlas CI agent supporting the creation of competitive landscapes, drug profiles, catalyst timelines and leadership briefings.

The company said these outputs can be generated in minutes rather than the days or weeks traditionally required for some competitive intelligence activities.

Kris Kaneta, chief product & innovation officer at Norstella, said: “There are two kinds of AI in this industry right now. One is built for demos. The other is built for decisions. Atlas is built for decisions.”

The platform is built around what Norstella calls its Clarity Test, an internal framework assessing whether an AI-generated answer is suitable for use in decision-making.

According to the company, Atlas outputs are assessed against four criteria: whether an answer is traceable to a source, whether it provides consistent results, whether it is relevant to the decision workflow and whether it is actionable.

The approach reflects a broader challenge facing generative and agentic AI in life sciences, where inaccurate or poorly sourced information can have significant consequences for drug development and commercial decisions.

By focusing on source traceability and contextual outputs, Norstella is attempting to position Atlas as a system for professional workflows rather than a general-purpose AI assistant.

Platform combines proprietary data with AI agents

Norstella said Atlas draws on data spanning clinical development, commercial activity, market access and real-world evidence.

The company’s data assets include information from Citeline, Evaluate, MMIT and Panalgo, with the platform designed to connect information relating to trials, drugs, companies, payers, plans and other events.

The company said this connected data foundation is intended to give Atlas access to context that can support more specialised outputs.

Atlas can also combine Norstella’s proprietary data with a customer’s internal intelligence, according to the company. This is intended to allow outputs to reflect both broader market information and information held within an individual organisation.

The approach could be particularly relevant to pharmaceutical and biotechnology companies, where decision-making often requires teams to combine external market intelligence with proprietary research, development and commercial information.

Norstella said its existing AI-enabled products and solutions, including Citeline SmartSolutions, Evaluate’s Ella and Omnium, MMIT’s Mosaic and Searchlight, and NorstellaLinQ, are already being used internally as part of the development and operation of its AI capabilities.

Future agents to target drug development workflows

Atlas CI is the first of a planned series of persona-specific agents that Norstella intends to develop for different functions across the biopharmaceutical industry.

Planned applications include business development and licensing, feasibility, portfolio strategy and protocol design.

The company said a shared platform architecture will provide common standards for security, sourcing, governance and transparency across the different agents.

Kaneta said: “Atlas is built for decisions. That takes two things nobody else has combined: a context layer that took years of expert curation to build, and agents designed around the person doing the work, not around the data underneath.”

The launch comes as pharmaceutical companies and technology providers increasingly explore how agentic AI could be applied to complex research and business processes.

While generative AI has already been adopted for tasks such as summarisation, drafting and information retrieval, agentic systems are being developed to perform more structured sequences of activities and produce outputs tailored to specific workflows.

For life sciences companies, potential applications span the drug development lifecycle, from early research and clinical trial planning through to commercial strategy and market access.

Norstella’s approach is to combine these capabilities with specialist data and human expertise, with the company emphasising that people remain involved in the decision-making process.

The initial launch of Atlas CI provides an early example of how persona-specific AI agents could be applied to pharmaceutical competitive intelligence. The planned expansion into feasibility, licensing, portfolio strategy and protocol design could extend the model across additional stages of drug development.

The effectiveness of such systems will ultimately depend on factors including data quality, transparency, reliability and their ability to deliver useful outputs within real-world workflows.

For now, Norstella’s Atlas launch highlights the direction of travel for AI in biopharma: from general-purpose tools that help users find and interpret information towards specialised agents designed to turn complex datasets into work that can support human decision-making.