H. pylori in the fine print: multi-agent detection in biopsies

In the field of digital pathology, one of the most complex challenges is the accurate extraction of clinical information from unstructured reports. A paradigmatic example is the detection of Helicobacter pylori in gastric biopsies, where data are often scattered across free-text and coded fields, with affirmations and negations requiring contextual interpretation. A recent study based on reports from Singapore demonstrated that a multi-agent system (nMAS) can achieve 98.6% accuracy in classifying key findings such as H. pylori positivity and associated gastritis. This approach not only improves scalability compared to manual review —which would consume over 80 hours per thousand reports— but also provides traceability by linking each decision to the source phrases. The lesson here transcends gastroenterology: when clinical data are hidden in the 'fine print' of reports, the combination of artificial intelligence for businesses and multi-agent architectures makes it possible to turn ambiguity into structured evidence. At Q2BSTUDIO, we know that these types of challenges are not exclusive to the healthcare sector. Many organizations face similar problems when extracting knowledge from legal, financial, or maintenance documents. That is why we develop custom applications that integrate AI agents specialized in natural language interpretation, capable of handling negations, contextual dependencies, and heterogeneous fields. Our AI agents are trained with real data and deployed in cloud environments (with AWS and Azure cloud services) to ensure scalability and cybersecurity. Additionally, we connect the results with business intelligence service platforms such as Power BI, allowing clinical or business teams to visualize metrics in real time without relying on spreadsheets. The key is not to simply copy manual workflows: the entire process must be redesigned, as the multi-agent system in the study did, where each agent handles a specific task (locating the biopsy, assessing the condition, identifying the pathogen) and then consolidates a verdict with evidence. This is exactly what we achieve with our custom software: transforming scattered data into auditable decisions. If your organization needs to extract value from unstructured reports —whether medical, technical, or commercial— a solution based on AI for businesses and intelligent agents can reduce review times from hours to minutes, freeing your team for higher-judgment tasks. At Q2BSTUDIO, we design and implement these systems, ensuring that each result is backed by the original source, as required by clinical practice or regulatory audit. Because when relevant information is in the fine print, well-applied technology ensures that no detail goes unnoticed.