Explore how AI is transforming regulatory intelligence in life sciences. Learn how UCB's quality and regulatory intelligence teams faced challenges and how they solved them with Infodesk's innovative solution which enables the observation, interpretation and activation of regulatory & quality intelligence.
Insights from a keynote by Barbara Bovy, Head of Quality, Regulatory Intelligence and QMS at UCB, and Brett Bonner, Chief Product & Technology Officer at Infodesk, presented at the RAPS Regulatory Intelligence Conference in Baltimore on March 3, 2026.
Regulatory and quality intelligence teams are doing the same core work: gathering signals, connecting the dots, interpreting what they mean, and deciding how to respond.
Two companies can see the same signal and reach different conclusions because they operate in different contexts. Even with complete information, the interpretation is situation-specific. Interpretation depends on where you are in the lifecycle – from early discovery to post-market vigilance.
Operating at the frontiers of medicine and process, for many teams, the week begins by checking health authority websites to see what has changed. Even after reviewing everything, the question remains: “what did we miss?” Teams working in regulatory and quality intelligence are flooded with updates. Time is limited, and the risk of missing the one update that truly matters is always present. We all have this anxiety: there is always the risk that leadership asks about an important update that slipped past your inbox.
The volume of warning letters is up. There are currently 424 guidances in draft stage across pharma & biologics for the FDA alone. Approximately 175 are in development right now, and this volume is exploding with 9,000 health authority published items per week*. That’s 9,000 updates each week that may carry compliance implications. Missing or misinterpreting the relevant update can lead to noncompliance.
*Infodesk source data
This reflects only the volume of documents published by health authorities – public health updates, guidance for industry, internal manuals made public, and news announcements.
The challenge is not only the volume. Complexity and applicability are increasing as well. Multiple layers of complexity now sit on top of this scale. Some signals are ambiguous. You also have some regulations that are very different and difficult to interpret. Understanding regulatory intent often requires significant effort. Small wording changes in regulatory text can significantly affect interpretation. In life sciences, success depends on scaling accurate interpretation of regulatory signals.
These challenges are common to all regulatory teams in life sciences organizations. Some organizations are stepping back to ask a fundamental question: is this level of information overload inevitable, or can the process and organizational structure be redesigned?
Leaders in regulatory and quality teams are addressing this challenge across functions. By having discussions with the different stakeholders across the organization they’re discovering a different organizational structure and regulatory intelligence process is needed. Cross-functional collaboration which unifies teams which had previously worked in silos is a fundamental key to solving the challenges of the regulatory landscape.
They are moving from manual monitoring to a more systematic approach to regulatory signal detection with a new process and system they can trust.
Today we are going beyond the traditional GxP to needing cross-functional expertise as a unified collective. This means when you receive a new signal or news about a new regulation you have a core team working on it together, who historically came from different departments, because you cannot address it in a single domain. This adds another layer of complexity on top of the growing volume.
Teams cannot simply scale headcount in proportion to the growth in volume, complexity, and applicability. A new solution is needed to address this: AI.
Large language models, when used within the right agent-based systems, can help support interpretation of ambiguous and complex regulatory text.
Adoption of AI in the pharmaceutical industry has been understandably cautious when balanced with the risks. This is partly due to the expectation that AI solutions must be near-perfect to be useful. Organizations are now recognizing that AI can start with lower-risk use cases and improve over time. For example, start with triaging intelligence. It’s a contained, lower-risk use case that allows teams to learn how AI performs in practice.
Regulatory authorities are increasingly demonstrating their own use of AI, which signals growing acceptance of the technology. In 2025, the FDA announced it was using its Elsa tool to support assessments and help prepare inspections. In June 2025, the FDA announced:
“Today marks the dawn of the AI era at the FDA with the release of Elsa, AI is no longer a distant promise but a dynamic force enhancing and optimizing the performance and potential of every employee,” said FDA Chief AI Officer Jeremy Walsh. “As we learn how employees are using the tool, our development team will be able to add capabilities and grow with the needs of employees and the agency. The agency is already using Elsa to accelerate clinical protocol reviews, shorten the time needed for scientific evaluations, and identify high-priority inspection targets. Elsa is a large language model–powered AI tool designed to assist with reading, writing, and summarizing. It can summarize adverse events to support safety profile assessments, perform faster label comparisons, and generate code to help develop databases for nonclinical applications. These are just a few examples of how Elsa will be used across the enterprise to improve operational efficiency.”
For companies such as UCB, this represents an important shift in how AI can be applied in regulated environments.
In software development, AI agents already assist with many tasks. The models are improving quickly. But turning intent into reliable software still requires significant interaction with the model and careful engineering. Chat tools can help interpret text, but they don’t remove the need to verify and validate the results. Confidence in the tools is growing, but outputs still need to be verified. Software development benefits from testing frameworks and guardrails, which make productivity gains possible. Today, the real bottlenecks are testing, verification, and deployment. The experience gained from applying agentic AI in our own development processes now informs how we can provide exert advice on AI to support customers.
Chat tools can help interpret regulatory documents. But when validating claims or identifying the exact differences between draft and final guidance, precision matters. Even a single word – such as ‘may’ versus ‘must’ – can change regulatory requirements. So for regulatory and quality intelligence a simple chat interface is not reliable enough.
Infodesk focuses on regulatory intelligence for life sciences. Using large language models and agentic systems, the platform helps scale human judgment when interpreting regulatory text. Infodesk combines large language models with structured agentic workflows to support reliable interpretation of regulatory text.
Regulatory intelligence activation involves three steps:
Continuous regulatory sensing identifies emerging signals across the regulatory landscape. Agents scan health authorities and track developments across guidance, enforcement, safety, and competitive activity.
Authority & Guidance
Enforcement & Safety
Competitive Intelligence
Emerging Signals
Agentic AI with humans in the loop helps teams interpret ambiguous regulatory text more consistently.
Guidance Analysis
Strategic Pathway Evaluation
Enforcement Patterns
Competitive Impact
Turning insight into regulatory decisions and outcomes.
Regulatory Strategy & Submissions
Post-Market & Organizational
Barbara Bovy, Head of QA at UCB, shared key insights from the company’s journey toward regulatory intelligence activation, in partnership with Infodesk.
UCB identified that their core challenge was structural and organizational, and could no longer rely solely on traditional GxPs. They recognised the need for a single, unifying platform for regulatory intelligence and quality intelligence for the entire company. Having one single platform has helped UCB to move away from the fragmented system they had which relied on e-mail and manual effort to bridge the silos.
Recognizing that humans are needed ‘in the loop’ is important. UCB recognized the need to first achieve alignment among people across all stakeholder functions before asking AI to support the process. It’s up to humans to define the elements and the processes, before AI can be trained to support those.
Make sure you start with sound and documented processes in your SOPs. Teams can start gradually and build confidence as AI becomes part of the workflow.
The pace of change in life sciences is accelerating. Device engineering and drug discovery are accelerating. We all hear about AlphaFold 3. What used to take a PhD candidate about five years to study a protein structure, now there’s over 300 million viable protein structures out there. There are cell models, there’s whole organism models, all of that is helping with the drug discovery process. But despite this approvals and the regulatory work stream are constraints. Infodesk helps regulatory teams move through this complexity and to remove the constraints with clearer intelligence and coordinated action.
Speak to us to learn more about how Infodesk can help you with regulatory and quality intelligence activation powered by AI.
“At UCB we recognized that regulatory intelligence only becomes value when it travels the full distance from signal to decision –and from decision to visible change. We knew we needed to move beyond pure monitoring towards “activation” of regulatory intelligence and turn raw information into decision-ready insight. Working with Infodesk has enabled us to move from monitoring to activation. For UCB, this partnership with Infodesk is helping us transform a growing volume of unstructured updates into a single, structured stream of intelligence that flows into our digital backbone and supports our Compliance Domains. Infodesk has been identified as the key digital partner to facilitate the regulatory intelligence journey. Reporting on Implementation, as defined by the E2E regulation Intelligence project, and subsequently applying Infodesk to real signals, real decisions and real changes – is where orchestration becomes culture.” Barbara Bovy – Head Quality Intelligence & QMS, UCB
“At UCB we recognized that regulatory intelligence only becomes value when it travels the full distance from signal to decision –and from decision to visible change. We knew we needed to move beyond pure monitoring towards “activation” of regulatory intelligence and turn raw information into decision-ready insight. Working with Infodesk has enabled us to move from monitoring to activation. For UCB, this partnership with Infodesk is helping us transform a growing volume of unstructured updates into a single, structured stream of intelligence that flows into our digital backbone and supports our Compliance Domains. Infodesk has been identified as the key digital partner to facilitate the regulatory intelligence journey. Reporting on Implementation, as defined by the E2E regulation Intelligence project, and subsequently applying Infodesk to real signals, real decisions and real changes – is where orchestration becomes culture.”
If you found this interesting look at these next…
Webinars
Read more
Blogs