Real-world validation
Our pilots
ARGENTIC is tested in seven real-world pilots across Europe to show that safe and sustainable AI works in practice.
Explore the pilots
Choose a sector to see what each pilot does and how ARGENTIC makes a difference.
Pilot 01 · Healthcare
Bedside AI that spots danger early
The challenge
Hospitals need to notice quickly when a patient's condition is getting worse, without putting sensitive health data at risk.
What ARGENTIC does
Small AI models run directly on medical devices at the patient's bedside, analysing heart activity, oxygen levels, heart rate and blood pressure in real time. They warn clinicians early about risks such as heart rhythm problems or internal bleeding, and explain why. Patient data never leaves the hospital.
- Real-time vital signs
- Explainable alerts
- Data stays in the hospital
Technical description
UC1: Bedside Agentic AI Monitoring for Proactive Hospital Care
Hospitals require rapid, secure analysis of patient vital signs to detect clinical deterioration early; however, conventional centralized cloud architectures introduce latency and data privacy risks. This pilot develops and validates the ARGENT-Care system, deploying lightweight Small Language Models (SLMs) directly on medical edge devices at the patient’s bedside. These models continuously process real-time data streams—including ECG, oxygen saturation, heart rate, and blood pressure—to predict risks of cardiovascular events (e.g., arrhythmias, atrial fibrillation) or internal hemorrhaging before standard clinical thresholds are breached. The system generates explainable alerts paired with interpretability scores (SHAP) on clinical dashboards, while model updates are performed via federated learning without primary patient data ever leaving hospital premises.
Pilot 02 · Smart Energy & Industry
Smarter energy use for factories and SMEs
The challenge
Industrial sites and small businesses want to cut energy costs, but cannot share sensitive operational data to do so.
What ARGENTIC does
AI agents added to the MangroviaIoT platform monitor energy use and make decisions on site, quickly and at lower cost. Digital twins test savings scenarios, such as better heating and cooling or smarter production schedules, and companies learn from each other without revealing their data.
- Energy monitoring
- Digital twins
- Shared learning, private data
Technical description
UC2: Mangrovia IoT (Smart Energy & Industry)
This pilot upgrades the MangroviaIoT industrial management platform into a sovereign, federated IoT layer integrated with ARGENTIC agents, enabling continuous energy monitoring and decentralized decision-making for industrial facilities and SMEs. Lightweight edge agents process sensor data locally, significantly reducing latency and operational energy costs. Concurrently, Digital Twins simulate energy efficiency scenarios—such as HVAC optimization and process scheduling—while federated learning empowers SMEs to benefit from cross-facility energy intelligence without exposing proprietary or commercial operational data.
Pilot 03 · Digital Marketing & E-commerce
Smarter, privacy-friendly advertising
The challenge
Online shops struggle to decide where to invest their advertising budget, especially now that AI is changing how people find products.
What ARGENTIC does
AI agents learn from a company's own data and from outside signals, such as weather and market trends, and recommend where to spend, explaining their reasoning. They simulate the return on investment before any money is spent and help brands stay visible in AI-powered search, in full compliance with GDPR and the EU AI Act.
- Explainable recommendations
- ROI simulations
- GDPR & AI Act compliant
Technical description
UC3: Digital Marketing Optimization (Digital Marketing & E-commerce)
European e-commerce businesses face an increasingly fragmented advertising landscape, compounded by generative AI fundamentally altering how consumers discover products. ARGENTIC introduces an agentic AI approach, moving away from static multi-touch attribution toward federated models that learn from both first-party data (CRM, web analytics) and external signals (weather patterns, market trends, mobility). AI agents deliver explainable channel allocation recommendations and Digital Twin ROI simulations under the supervision of marketing teams (human-in-the-loop), maintaining full GDPR and EU AI Act compliance. A federated marketing knowledge graph enables cross-actor collaboration without raw data exchange. Furthermore, the agents optimize brand visibility for GenAI-driven search engines (AI-first positioning) to sustain enterprise competitiveness.
Pilot 04 · Culture & Tourism
Personal AR/VR guides for cultural heritage
The challenge
Visitors want tours that match their own interests, while personal information such as their voice must stay private.
What ARGENTIC does
AI agents in AR/VR headsets create live, personalised narrations based on each visitor's interests, pace and available time. Voice and viewing preferences are processed on the headset itself, and museums and heritage sites keep full control of their digital collections.
- Immersive AR/VR
- Personalised tours
- Privacy on the device
Technical description
UC4: Immersive AR/VR Cultural Heritage Personalized Guide (Culture & Tourism)
Developing next-generation immersive AR/VR experiences, this pilot deploys decentralized, edge-level AI agents to deliver real-time, explainable, and personalized guided tours. By coupling Large Language Models (LLMs) with federated cultural heritage knowledge bases and contextual cues (visitor location, pace, engagement), agents dynamically craft narrations tailored to each visitor’s interests and available time. Highly sensitive user inputs, such as voice prompts and viewing preferences, are processed on-device (headset edge) to guarantee privacy, while cultural institutions preserve complete governance and intellectual property over their digital assets.
Pilot 05 · Precision Agriculture
Smarter strawberry growing in Southern Europe
The challenge
Strawberry production in Southern Europe depends heavily on a single imported variety, which makes growers vulnerable.
What ARGENTIC does
Sensors and local AI monitor greenhouse climate, soil and irrigation in real time. Digital twins simulate droughts, salinity and heatwaves to develop growing methods that adapt to changing conditions, and nurseries share know-how without exchanging business data. The aim is to cut reliance on the imported variety by up to 50%.
- Greenhouse sensors
- Climate simulations
- Up to 50% less dependence
Technical description
UC5: AI-Driven Strawberry Breeding and Cultivation Optimisation (Precision Agriculture)
Focused on reducing South European strawberry production’s reliance on the imported Fortuna variety by up to 50%, this pilot deploys IoT sensor networks and edge AI to monitor greenhouse microclimates, substrate pH/EC, and irrigation/fertigation demands in real time. Digital Twins simulate critical stress scenarios (drought, salinity, heatwaves) to establish dynamic, adaptive cultivation protocols. Through federated learning, breeding insights and agronomic practices are shared across nursery networks without exchanging sensitive business data, strengthening European agricultural sovereignty and sustainability.
Pilot 06 · Human Resources & HR Tech
Fair and transparent AI-assisted hiring
The challenge
Hiring needs to become faster and more efficient, without compromising fairness, transparency or candidates' privacy.
What ARGENTIC does
AI agents in the Workable hiring platform guide new candidates, adapt interview questions to each person with strict fairness checks, and help detect fraud. People always make the final hiring decisions, and digital twins test the system thoroughly before it is used for real.
- Bias checks
- Humans decide
- Fraud detection
Technical description
UC6: Recruitment with Agentic AI: Efficiency, Oversight & Fairness (Human Resources & HR Tech)
Integrating agentic AI into the Workable hiring platform, this pilot optimizes recruitment pipelines while ensuring transparency, privacy, and bias mitigation. Features include conversational onboarding agents, adaptive video interview screeners that adjust questions dynamically according to candidate qualifications with strict fairness filters, and fraud detection mechanisms. A rigorous human-in-the-loop framework ensures that critical hiring decisions remain strictly under human supervision, supported by Digital Twins that simulate recruitment pathways to benchmark agentic reliability before deployment.
Pilot 07 · Adolescent Mental Health
MindBuddy: mental health support for teenagers
The challenge
Support for teenagers' mental health often stops at the clinic door, while everyday life happens at home and at school.
What ARGENTIC does
MindBuddy connects hospital care with daily life, through friendly social robots in hospitals and tablets at home. Local AI recognises emotional signals from voice, facial expressions and the body, and offers real-time exercises to help with mood and prevent crises. Clinicians always stay in charge and receive clear, easy-to-read summaries.
- Social robots
- AI that runs locally
- Clinicians in charge
Technical description
UC7: MindBuddy – Adolescent Mental Health Support (Adolescent Mental Health)
MindBuddy is a comprehensive digital and robotic support ecosystem designed to bridge specialized clinical environments with adolescents’ daily routines at home and school. Combining hospital-based social robots and in-home tablets powered by localized LLMs, the platform uses multimodal sensing (voice acoustics, facial cues, physiological signals) to gauge emotional states and offer real-time mood regulation and crisis prevention exercises. Human-in-the-loop governance guarantees that all therapeutic trajectories remain under clinician oversight, with practitioners receiving explainable, FHIR-compliant dashboard summaries.