| Q1Operator | CSL Plasma mobile app + iGive Rewards loyaltyDonor engagement | ✅ | Digital donor registration, appointment scheduling, loyalty programme management, and donor status tracking. Standard structured workflow assistance across 300+ US centres. 3.8M downloads by FY2024. App executes structured appointment and loyalty workflows; donors interact. No AI personalisation engine evidenced beyond standard loyalty tier rules.As of April 2026 |
| Q2Tactician | NLP for MedDRA auto-coding (Linguamatics)Pharmacovigilance | ✅ | NLP workflow for adverse event (AE) case processing auto-codes AE reports to MedDRA terminology. Doubled the auto-coding rate from 30% to over 60% of cases, deployed in post-market safety case processing. NLP generates auto-coded MedDRA assignments; human medical coders review and validate all cases, especially those requiring clinical judgment beyond exact lexical match.As of April 2026 |
| Q2Tactician | AI safety signal detection and NLP for PV signalsPharmacovigilance | 🔶 | NLP analysis of real-world data — call centre feeds, emails, regulatory reports, literature — to surface potential safety signals; AI identifies patterns and predicts risk signals. CSL Behring cited in Nature's 2026 industry analysis as employing this capability. NLP flags signals and extracts context; human pharmacovigilance physicians conduct benefit-risk assessment and determine reportability and regulatory action.As of April 2026 |
| Q2Tactician | Real-world evidence (RWE) analyticsSafety sciences / commercial | ✅ | Mining physician notes, patient charts, registries, and EHR data for evidence generation; predicts patient subgroups at risk of adverse events and demonstrates product effectiveness to regulators and payers. Led by Haley Kaplowitz, EVP Global Head Safety Sciences. Analytics models process structured and unstructured clinical data; human scientists and regulatory strategists interpret outputs and compose regulatory submissions and market access evidence packages.As of April 2026 |
| Q2Tactician | Donor Health Vault (AWS — inferred)Donor health data | ◻️ | Data infrastructure on AWS Digital Core, named as a critical business initiative alongside R&D and Deviation AI. Specific functionality is not publicly disclosed. Classification is inferred because donor health data analytics in a regulated health context most naturally supports augmented human decision-making.As of April 2026 |
| Q3Automator | Rika Plasma Donation SystemPlasma collection | ✅ | Machine-led plasma collection procedure at all US CSL Plasma centres. Completes an average donation in under 35 minutes, maintains no more than 200 ml of extracorporeal blood, and provides an advanced touchscreen HMI that guides operators through alarms plus real-time data analytics for device maintenance. Machine executes the FDA-cleared collection protocol; operator intervenes on alarms per HMI-guided resolution steps.As of April 2026 |
| Q3Automator | iNomi individualised nomogramPlasma collection | ✅ | Calculates donor-specific collection volume from height, weight, and haematocrit (28.5% of total plasma volume, up to 1,000 ml). FDA 510(k) cleared in 2024; deployed across all US and Puerto Rico CSL Plasma centres in FY25, delivering an average 10% increase in plasma volume per donation. The Rika system calculates and applies the iNomi volume target; the donor provides biometric inputs and the machine executes with no operator discretion over the calculation.As of April 2026 |
| Q3Automator | Project Aurora — manufacturing robotics (Broadmeadows Facility F)Manufacturing | ✅ | Robotic plasma pooling, cleaning, powder handling, and waste disposal; automated plasma debottling/debagging; COP robotic filter press machines; an AGV warehouse system; Siemens PCS-7 process control; and Körber PAS-X MES. Nine-fold capacity expansion, ISPE Facility of the Year (Pharma 4.0) 2025. Robots and automated systems execute defined manufacturing steps within GMP protocols; operators oversee, manage deviations, and authorise batch releases. No AI adaptive control of production parameters was deployed as of September 2025.As of April 2026 |
| Q3Automator | Pharmacovigilance RPA (Oracle Argus)Pharmacovigilance | ✅ | Robotic process automation handles non-value-added PV tasks: data entry, case routing, document retrieval, and standard case processing workflows, freeing physicians to focus on benefit-risk management. Oracle Argus is the deployed safety case management platform. RPA handles structured repetitive tasks within defined workflow rules; physicians conduct all clinical evaluation, benefit-risk assessment, and regulatory reporting decisions.As of April 2026 |
| Q3Automator | AWS TMP — agentic AI wave planning (IT migration)IT infrastructure | ✅ | Agentic AI (AWS Transform for VMware) applied to IT migration planning automated wave planning for 1,072 applications across 29 data centres, saving 10.5 weeks of manual effort and accelerating application discovery 12x. Delivered an estimated 30% operational cost savings from cloud migration. AWS Transform agents analysed dependencies and generated migration wave plans; IT architects reviewed and approved plans; human teams executed migrations.As of April 2026 |
| Q3Automator | RISE with SAP S/4HANA (Project Unify)Enterprise systems | 🔶 | Consolidating three separate SAP ERP instances into one unified RISE with SAP S/4HANA deployment on AWS, establishing enterprise-wide data standards and governance. Two of three deployments were complete by December 2025. The ERP automates structured enterprise processes — procurement, finance, HR, batch tracking — while IT and business users manage configuration and exceptions.As of April 2026 |
| Q4Weaver | PlasmIX mixed reality + digital twin (Bern)Manufacturing | 🔶 | Microsoft HoloLens glasses with a smart AR cloud digital twin display real-time production data on physical fractionation equipment in the cleanroom, enabling a hands-free operator experience with faster decisions and error avoidance. Won Gold at the Best of Swiss App Awards 2018. The human operator composes all process decisions; the AR system supplies contextual real-time data on the physical object in the operator's field of view without recommending or executing actions.As of April 2026 |
| Q4Weaver | Veeva CTMS + operational data lakeClinical operations | 🔶 | A unified clinical trial management system across CSL Behring and Seqirus, with an operational data lake for data capture, reporting, and predictive analytics that supports trial execution optimisation. Human clinical operations teams use the platform to compose and execute trial management decisions across complex, evolving clinical workflows; analytics are decision-support, not autonomous.As of April 2026 |
| Q5Partner | R&D bioinformatics + omics AI (Global Research Data Science)Drug discovery | 🔶 | Multi-omics — genomics, transcriptomics, proteomics — ML models and a knowledge graph representation of biomedical data generate disease hypotheses validated in the wet lab, including a 'reverse translational medicine' approach that feeds clinical non-responder insights back to the bench for next-generation therapy design. Led by Monther Alhamdoosh, Director of Bioinformatics and AI. AI generates hypotheses and surfaces patterns across massive multi-dimensional omics datasets; human scientists evaluate biological plausibility, design validation experiments, and govern which hypotheses advance.As of April 2026 |
| Q5Partner | Deviation AI (AWS — clinical trial analytics)Clinical operations | 🔶 | AI-powered analytics find deviations and outliers in clinical trial outcomes, enabling faster evaluation and adjustments to trial protocols. Deployed on the AWS Data and Analytics Platform across three Regions, and listed as a critical business initiative alongside R&D and Donor Health Vault. Human clinical teams act on Deviation AI outputs: when deviations are identified, clinical scientists evaluate root causes and determine protocol adjustments.As of April 2026 |
| Q5Partner | Project Aurora — SIMIT digital twin (Broadmeadows)Manufacturing | 🔶 | A Siemens SIMIT virtual replica of Facility F's automation systems simulates and validates process changes before live deployment, tests scenarios, identifies bottlenecks, validates control logic without interrupting production, and supports operator training in simulated environments. Engineers use the digital twin to iteratively explore process options and validate changes; human process engineers compose all decisions about what changes to implement in the physical plant.As of April 2026 |