AI-industrie.png

Smart Manufacturing & Materials-iStock-1135159658-BEWERKT.jpg

<p><strong>Drug shortages are a growing problem in the Netherlands. For patients, this means uncertainty; for pharmacists, extra work; and for manufacturers, complex decisions. The Nijmegen-based AI company RepliKate, in collaboration with a pharmaceutical company, developed a solution that uses AI agents to continuously monitor thousands of medications. The goal: to identify shortages earlier and provide substantiated evidence of when production is needed.</strong></p><p><img src="https://cms.aihub-oost.nl/uploads/Repli_Kate_019b0b27dd.jpg?token=DY59zjKy5dJfssWxHCqW9D6KTdJU-eDDXYVOmNDsuyc&amp;expires=1777471984" alt="RepliKate.jpg" srcset="https://cms.aihub-oost.nl/uploads/thumbnail_Repli_Kate_019b0b27dd.jpg?token=rOY87pZ8W5gkP7sTO2wBE4FbJyKfoB0FGogKBLZ0GEU&amp;expires=1777471984 245w,https://cms.aihub-oost.nl/uploads/small_Repli_Kate_019b0b27dd.jpg?token=OGMexfF4R2fXAFOAw_FPbbI2NW6YBe7GZrL36KwOcLU&amp;expires=1777471984 500w,https://cms.aihub-oost.nl/uploads/medium_Repli_Kate_019b0b27dd.jpg?token=xadzLTgnzn3DP0j4Z2zn8QtYrWxL0hMYsIt5X8ETMjY&amp;expires=1777471984 750w,https://cms.aihub-oost.nl/uploads/large_Repli_Kate_019b0b27dd.jpg?token=p_JXAzKQBnlOWDangP-7LtT7HhTJznQlaOLTDmgppKo&amp;expires=1777471984 1000w," sizes="100vw" width="2048" height="1153"></p><h2>From Manual Screening to Smart Detection </h2><p>The pharmaceutical company’s question was clear: “How can we work more efficiently with AI?” In joint workshops, AI specialists and domain experts pooled their knowledge. A specific issue quickly emerged.</p><p>For certain medications, production may only take place when a shortage can be demonstrated. In the Netherlands, we have approximately 20,000 registered medications. To determine whether a shortage is imminent, sales figures, the number of suppliers, and possible alternatives must be taken into account. For example, a 500 mg tablet can be replaced by a divisible 1,000 mg variant. Such nuances make all the difference.</p><p>The initial screening of all these medications is carried out by highly trained professionals, but it is largely repetitive. “That’s typically an area where AI can add value,” says Robin van Breukelen of the AI company RepliKate. “It eliminates the repetitive work and makes the assessment more comprehensive than when you rely primarily on experience or ‘gut feeling.’”</p><h2>Two AI Agents in Practice </h2><p>The collaboration between the pharmaceutical company and RepliKate resulted in two applications. First, RepliKate developed an AI platform with a chat interface. Employees can ask questions about a specific medication, after which the AI automatically combines multiple data sources into a single comprehensive answer. Instead of manually searching through different systems, they gain immediate insight.</p><p>In addition, the team built an algorithm that weighs various indicators against each other and predicts which medications are at the highest risk of a shortage. The AI agent continuously scans the entire list of 20,000 medications and flags the records that warrant further analysis. This allows professionals to focus their attention on the medications where the risk is greatest, while monitoring is conducted more broadly and consistently.</p><h2>The final 20 percent makes all the difference </h2><p>Technically, the combination of diverse data sources proved particularly challenging. Different formats and definitions required thorough data cleansing. It’s striking that the first 80 percent of an AI solution can be implemented relatively quickly. Generic models and standard functionalities take you a long way. But the real value lies in the final 20 percent: adding specific domain knowledge, correctly handling exceptions, and ensuring reliability.</p><blockquote><p>“The devil is in the details,” says Van Breukelen. “It’s about source citations, traceability, and preventing hallucinations. You shouldn’t underestimate that.”</p></blockquote><h2>Collaboration as a Key Success Factor </h2><p>The collaboration between AI specialists and pharmaceutical experts was crucial. RepliKate contributed technological expertise, while the pharmaceutical company provided in-depth knowledge of regulations and drug registration. “You can’t just throw a project like this over the fence,” says Van Breukelen. “It’s really about exchanging knowledge. The better both parties understand each other’s world, the stronger the solution becomes.”</p><p>Especially in a regulated market like pharmaceuticals, transparency and explainability are prerequisites. Citing sources and understanding the origin of data are essential to maintaining trust. With this application, RepliKate demonstrates how AI agents not only increase efficiency but also contribute to addressing a societal challenge: reducing drug shortages and making smarter use of scarce expertise in the healthcare supply chain.</p><h2>More Opportunities for AI Agents in the Supply Chain </h2><p>Although this application focuses on shortage situations, RepliKate sees broader opportunities for AI agents in the pharmaceutical supply chain. Examples include searching through complex registration dossiers, combining external market data, or supporting policy decisions.</p><p>For AI companies looking to enter this market, Van Breukelen has a clear message: don’t underestimate the value of domain knowledge. “As a tech person, you tend to take big leaps and think, ‘It must work this way.’ But in pharma, things are very precise. You really have to understand how processes and definitions work.”</p><h2>Practical tips for AI companies looking to collaborate with the pharmaceutical or biotech industries </h2><p><strong>1. Invest in domain knowledge early on</strong>.<br>Technology alone isn’t enough. Understand regulations, definitions, and exceptions.</p><p><strong>2. Collaborate iteratively</strong>—<br>don’t just hand off projects; ensure continuous alignment between tech and content.</p><p><strong>3. Ensure traceability and source attribution</strong><br>Especially in regulated markets, explainability isn’t a luxury—it’s a prerequisite.</p><p><strong>4. Pay attention to data quality and cleansing.</strong><br>Different data sources require thorough preparation.</p><p><strong>5. Focus on the final 20 percent.</strong><br>The real value lies in refinement, not in the generic foundation. </p>

RepliKate helps pharmaceutical companies identify drug shortages earlier using AI agents

These entrepreneurs have paved the way for you

Meet our frontrunners

RepliKate.jpg

<p>Drug shortages are a growing problem in the Netherlands. For patients, this means uncertainty; for pharmacists, extra work; and for manufacturers, complex decisions. The Nijmegen-based AI company RepliKate, in collaboration with a pharmaceutical company, developed a solution that uses AI agents to continuously monitor thousands of medications. The goal: to identify shortages earlier and provide substantiated evidence of when production is needed.&nbsp;</p>