Top 10 AI Prompts and Use Cases and in the Healthcare Industry in Spain

Spain's Sistema Nacional de Salud (SNS) delivers near‑universal, tax‑funded care through 17 autonomous regions, so AI matters not as a flashy add‑on but as a practical tool to stitch fragmented data, speed diagnoses and protect a system that covers roughly 99.5% of residents and ranks among the EU's healthiest populations (see WHO's Spain health system review (2024)).

With public spending shouldering about 71–72% of costs and rising pressure from chronic disease, rural primary‑care shortages and the need to curb low‑value procedures, targeted AI - predictive analytics for early detection, clinical decision support and administrative automation - can reduce avoidable admissions and free clinicians for front‑line care.

For professionals who want to turn these use cases into workplace impact, the AI Essentials for Work bootcamp teaches promptcraft and practical AI skills to build useful, ethically aware tools for Spain's decentralized health landscape.

Selection for the Top 10 emphasised Spain‑ready, beginner‑friendly prompts and use cases and was anchored in the FUTURE‑AI international consensus guideline - a 117‑expert framework built around six guiding principles and 30 best practices to make clinical AI trustworthy and deployable (FUTURE‑AI consensus guideline).

Priority filters included demonstrable fairness (bias testing across subgroups), universality with local clinical validity (external validation and recalibration so tools work across Spain's autonomous regions), traceability and audit trails, clinician‑centered usability, robustness to real‑world variation, and explainability for safe human oversight.

Practical requirements also demanded documented evaluation plans, stakeholder engagement, and early regulatory alignment for high‑risk healthcare AI. To keep the list actionable for beginners, the methodology favoured use cases that map to SNHS workflows, can be validated via federated approaches, and include clear deployment steps (monitoring, training and logging) - turning abstract promise into a concrete, safety‑minded playbook for adoption.

FUTURE‑AI Principles	Focus
Fairness	Bias identification and mitigation
Universality	External validation & local fit
Traceability	Documentation, logging, audits
Usability	Human‑AI interfaces & training
Robustness	Stress testing & representative data
Explainability	Clinically meaningful explanations

Owkin's CE‑marked MSIntuit® CRC brings digital pathology into routine MSI pre‑screening by analysing digitized H&E/HES slides to triage microsatellite stable (MSS) cases and spare confirmatory IHC/PCR - a practical fit for Spanish pathology labs that have adopted whole‑slide imaging and IMS workflows.

The tool's published validation shows real‑world impact (almost 50% rule‑out on biopsies and up to 70% on resections), very high sensitivity (≈95–96%) and a 99% NPV at a 10% MSI prevalence, meaning many patients avoid extra testing while precious tissue and auto‑stainer time are conserved for high‑value use.

MSIntuit® CRC v2 expands biopsy support and is being rolled out in research platforms with partners; Spanish centers evaluating digital readiness can leverage these integrations to accelerate therapeutic decision pathways and reduce lab bottlenecks.

Learn more on Owkin's MSIntuit® CRC product page, the v2 press release, and the Nature Communications validation for the evidence base behind the claims.

Metric	Result / Note
Rule‑out	≈50% (biopsies), ≈70% (surgical resections)
Sensitivity	≈95% (biopsies), 96% (resections)
Negative predictive value (NPV)	99% (reported, 10% MSI prevalence)
Regulatory status	CE‑marked (MSIntuit® CRC v1, 2022); v2 under development/RUO
Intended use	Adults with primary colorectal adenocarcinoma; workflow‑agnostic integration with WSI/IMS

“Collaboration with MSD allows us to bring our cutting‑edge AI technology to a broader audience, ultimately benefiting more patients and healthcare systems.” - Meriem Sefta, PhD, Chief Diagnostics Officer, Owkin

Owkin's four‑year collaboration with IRYCIS at Ramón y Cajal is a Spain‑focused example of precision oncology in action: by building an AI‑ready, GDPR‑compliant local database and IT stack, the partnership aims to use multimodal models to spot which early‑stage prostate cancer patients - remember, roughly one in eight men will be diagnosed in their lifetime - are likely to fail primary curative treatment and might benefit from earlier, personalised interventions; the first project, funded with Janssen support, is explicitly designed to keep data local and privacy‑preserving while unlocking clinically useful signals.

This Spanish work dovetails with European federated efforts such as the OPTIMA programme - using Owkin's federated learning approach to extract insights without pooling data - to create interoperable evidence platforms for prostate, breast and lung cancer, and with larger multimodal mapping initiatives like ATLANTIS that surface harmonised datasets across countries (including Spain) to accelerate biomarker and treatment‑selection models.

For SNHS clinicians and data teams, these projects show a practical path from curated, local data governance to AI tools that can be validated, audited and integrated into oncology pathways.

Initiative	Focus	Note
Owkin–IRYCIS precision oncology collaboration in Spain	Early‑stage prostate cancer treatment selection	4‑year partnership (from Sep 28, 2023); first project funded by Janssen; GDPR & LO 3/2018 compliant
OPTIMA (IMI) federated AI research project for prostate, breast and lung cancer	Federated AI for prostate, breast, lung cancer	€21.3M public‑private programme; federated learning to avoid data pooling
ATLANTIS	Multimodal patient data mapping	Launched Feb 27, 2025; spans 11 therapeutic areas across 7 countries including Spain; completion expected May 2025

“Our research objectives are always driven by impacting patients' lives, and in this sense, the current project is going one step forward in the personalized medicine field.” - Fernando López Campos, MD, PhD, MSc, Clinical Oncologist, Ramón y Cajal Hospital

Almirall's Barcelona headquarters is putting Spain squarely on the map for AI-accelerated dermatology R&D by deepening a partnership with Absci that has already produced AI-designed, functional antibody leads and now targets a second “difficult‑to‑drug” dermatology molecule; the collaboration pairs Almirall's clinical development expertise with Absci's Integrated Drug Creation™ generative‑AI plus wet‑lab loop to shorten design‑to‑candidate cycles and aim for faster time‑to‑clinic, supported locally by high‑performance computing links with the Barcelona Supercomputing Center.

The deal - launched in November 2023 and expanded in mid‑2025 - signals a practical, Spain‑relevant route for biotech AI to tackle chronic skin diseases, mobilise domestic scientific infrastructure, and attract sizable investment (Absci is eligible for up to approximately $650 million across the two programmes).

For concise coverage see Almirall's announcement and PharmTech's report on the second target, both of which outline roles, timelines and the strategic move toward first‑in‑class biologics for patients with severe dermatological conditions.

Item	Detail
Partnership launch	November 2023
Expansion / 2nd target	Announced August 2025
Platform	Absci Integrated Drug Creation™ (generative AI + wet lab)
Financial terms	Up to approximately $650 million in milestones & royalties
Almirall HQ	Barcelona, Spain

“Using advanced AI capabilities to design therapeutic candidates against historically challenging disease targets is a highly promising approach and Absci´s de-novo AI platform capabilities have already demonstrated early success. We are pleased to expand our collaboration as we continue to harness AI to help us develop innovative treatments for patients living with severe skin conditions.” - Karl Ziegelbauer, PhD, Chief Scientific Officer, Almirall

IMPaCT-style population health work in Spain hinges on practical cohort discovery and scalable analytics: platforms like Dataiku RWD Cohort Discovery cohort analytics platform make reusable, OMOP‑mapped cohorts and dashboards so teams spend less time stitching SQL pipelines and more time testing interventions, while HDR UK's Cohort Discovery shows how researchers can

search for women aged 18–30 with asthma who do not smoke

Complementing that, the Johns Hopkins ACG System population health risk stratification brings mature risk stratification and social‑determinant inputs that help turn cohort lists into actionable priorities for prevention, resource planning and equity audits.

For Spain's regional health networks an IMPaCT approach therefore combines interoperable cohorts, federated access and validated risk models - so teams can spot high‑impact groups quickly (a tiny, overlooked cohort can drive outsized hospital use) and plan targeted, measurable interventions.

Capability	Source / Benefit
Cohort discovery & OMOP mapping	Dataiku & HDR UK - reusable, interoperable cohorts
Risk stratification & SDOH	Johns Hopkins ACG - 350 dimensions for population risk
Federated, privacy‑preserving search	HDR UK example - query multiple datasets without moving data

“AI may be a useful tool, but alone it will never achieve all that.” - Jonathan Weiner, Johns Hopkins Bloomberg School of Public Health

Hispabot‑COVID19 and AsistenciaCOVID‑19 sit squarely in the practical, Spanish-facing corner of conversational AI: these virtual assistants channel the same symptom‑triage, appointment‑reminder and multilingual support roles highlighted in The Medical Futurist's roundup of top healthcare chatbots, offering a scalable first contact point that can keep low‑acuity queries out of clinics and free clinicians for higher‑value care (Medical Futurist roundup of top healthcare chatbots).

Spanish research shows this approach reaches beyond triage: the JMIR‑published Vickybot work demonstrates chatbots can screen anxiety and depressive symptoms in primary care and among health workers, making conversational agents a real tool for mental‑health follow‑up (Vickybot study on anxiety and depressive symptoms (JMIR 2023)).

Yet real‑world evidence warns of design risks: a large deployment study found roughly 35.6% of consultations dropped before diagnosis and nontherapeutic (gaming) interactions in about 8% of completed sessions, underscoring the need for clear onboarding, privacy safeguards and actionable outputs if Spanish virtual assistants are to earn trust and sustained use (real‑world chatbot evaluation (JMIR 2021)).

A memorable takeaway: conversational AI can unclog workflows, but unless users stay engaged beyond the first few rounds, its promise stays theoretical rather than practical.

Capability / Finding	Source / Metric
Common uses	Symptom triage, reminders, multilingual support (Medical Futurist; Simbo.ai)
Mental‑health screening	Vickybot: useful for anxiety, depressive symptoms and burnout screening (JMIR 2023)
Real‑world engagement risks	DoctorBot study: 35.6% consultations dropped; 8.03% nontherapeutic use (JMIR 2021)

Clinical decision support and workflow automation are where AI can move from pilot to everyday benefit in Spain's SNS - practical examples include the KNOWBED system, a JMIR study that developed and evaluated a bedside clinical decision support tool to integrate scientific knowledge into point‑of‑care decisions (KNOWBED bedside clinical decision support, JMIR 2021), and a 2025 systematic review showing that secondary use of electronic health records can reliably support prediction, detection and treatment‑planning tasks that feed CDS engines (Systematic review: secondary use of EHRs for prediction and treatment planning, BMC Medical Informatics 2025).

For the SNHS, the practical “how” matters: pairing bedside CDS with privacy‑aware, locally validated data strategies - such as federated networks described in our Spain guide - lets regions reuse data without centralising it, maintain audit trails, and automate routine workflows so clinicians see one succinct, evidence‑linked recommendation rather than hunting through multiple sources.

Source	Relevance for SNHS CDS
KNOWBED bedside clinical decision support (JMIR 2021)	Bedside CDS to integrate scientific knowledge into clinical workflows
Systematic review on secondary use of electronic health records for CDS (BMC Medical Informatics 2025)	Evidence that secondary use of EHRs can enable prediction, detection and treatment recommendations

LEIA's stated aim -

the defence, projection and correct use of the Spanish language in the field of Artificial Intelligence and current technologies

- puts language quality front and centre for any Spanish clinical NLP effort, and that mission is reflected in the work of Telefónica's Chief AI & Data Strategist, Dr. Richard Benjamins (Telefónica Tech profile for Dr. Richard Benjamins).

Accurate, standardised Spanish clinical text is what lets NLP reliably extract diagnoses, map to billing and coding systems, and produce clean inputs for federated research networks so models trained locally can be meaningfully compared across Spain's autonomous regions; for teams building SNHS‑ready pipelines this is precisely why federated approaches and language hygiene go hand‑in‑hand.

The practical payoff is memorable: invest once in linguistic standards and tooling and free‑text discharge notes stop being a

mystery puzzle

for downstream analytics - turning messy prose into machine‑actionable data that speeds coding, reduces downstream rework and makes population‑level predictions usable across all 19 regional systems.

MAPFRE's pragmatic AI playbook offers a concise blueprint for Spanish healthcare administrators who want to cut paperwork and protect revenue: automated claims engines that read policies, photos and documents to settle simple cases in seconds (see the MAPFRE–Shift claims automation pilot), plus virtual agents and internal assistants such as MIA GPT and customer bots that handle high volumes of enquiries and routine transactions - capabilities MAPFRE already applies to insurers and is beginning to adapt for digital physiotherapy and document automation.

For the SNHS, those same building blocks translate into faster prior‑authorization, automated coding and billing suggestions, streamlined patient intake, and digital rehab follow‑up that keeps mild cases out of busy clinics; the practical payoff is clear and measurable - faster turnarounds, fewer manual denials, and freed clinician time for higher‑value care rather than chasing forms.

Metric	Value (reported)	Source
AI use cases identified	Over 200	MAPFRE artificial intelligence overview
AI projects in development	More than 90	MAPFRE artificial intelligence overview
AI‑driven initiatives implemented	115+ globally	Reinsurance News: Mapfre introduces ethical guidelines for AI development
Virtual assistants handling share of transactions	40%+ (AI‑powered)	Reinsurance News: Mapfre introduces ethical guidelines for AI development

“AI must serve people - not replace them.” - Maribel Solanas, MAPFRE Group Chief Data Officer

Spain's MareNostrum 5 at the Barcelona Supercomputing Center (BSC) is turning high‑performance computing into a practical tool for chronic disease management: under the AI‑SPRINT personalised healthcare use case BSC pairs MareNostrum's raw power with an edge‑to‑cloud architecture to ingest continuous heart‑rate and rhythm streams from smartbands (Bluetooth to smartphones), combine them with lifestyle and biochemical data, and run GDPR‑aware, federated models that flag stroke risk and atrial fibrillation early - so a tiny irregular heartbeat picked up during a morning walk can trigger a targeted alert rather than a costly hospital visit.

The setup uses orchestration tools (COMPSs, OSCAR), secure enclaves (SCONE) and local storage syncing to simulate ‘hospital' edge nodes while protecting data, and even embeds device algorithms (e.g., the RITHMI AF detector) to maximise signal quality.

For Spanish health services this means scalable monitoring that empowers patients, prioritises high‑risk cohorts and helps regions use MareNostrum's compute to turn wearable streams into actionable prevention at population scale - literally putting a supercomputer housed in a Barcelona church to work on people's heartbeats in real time.

Item	Detail
Lead organisation	Barcelona Supercomputing Center - MareNostrum 5 (HPC)
Use case focus	Stroke risk assessment & prevention using wearables (AI‑SPRINT)
Key tech	Edge‑to‑cloud orchestration, federated learning, secure enclaves, smartbands + phones

“Catalonia takes an unprecedented step forward with the MareNostrum 5, one of the major objectives of European strategic autonomy and Catalonia's economic and social future” - Pere Aragonès, President of the Government of Catalonia

EXSCALATE4COV and Spain's DataCOVID together exemplify two complementary AI responses to a pandemic: EXSCALATE4COV paired supercomputing with AI to run one of the largest virtual‑screening experiments - a Europe‑wide public‑private consortium (including the Barcelona Supercomputing Center) that screened over 70 billion molecules and evaluated more than 1,000 billion interactions in a single, 60‑hour campaign, producing reusable resources such as the community MEDIATE portal for follow‑up testing - see the EXSCALATE4COV virtual screening project and the MEDIATE molecular docking portal for follow-up testing; meanwhile Spain's DataCOVID used big‑data analysis of population movement to inform regional epidemiology and resource planning as described in the national AI strategy, showing how AI can guide where beds, tests and outreach are needed most (Spain national AI strategy report on DataCOVID).

The practical takeaway for the SNHS is clear: combine federated, HPC‑backed in‑silico triage with population‑level analytics to cut months from candidate discovery and make scarce clinical capacity stretch farther when every hour counts.

Metric	Value / Note
Molecules screened (in silico)	>70,000,000,000
Total interactions evaluated	>1,000,000,000,000 (≈1 trillion)
Notable compute	Simultaneous use of Marconi100 & HPC5 (~81 PFLOPS); BSC participated
Community resources	MEDIATE portal (molecular docking at home) for follow‑up testing
Spain epidemiology tool	DataCOVID - big‑data mobility analysis to support regional decisions

The practical takeaway for beginners in Spain is straightforward: marry privacy‑first architectures with skills and small, measurable pilots so AI supports the SNHS rather than complicating it - start by aligning projects with GDPR‑compatible approaches like federated learning (the EDPS/AEPD TechDispatch highlights FL as compatible with GDPR and urges production‑grade platforms for secure, auditable use; see the Sherpa.ai summary), adopt Spain‑ready federated networks to span the 19 regional systems (our guide on federated data networks explains the how), and focus on high‑value, low‑risk wins such as predictive analytics for early detection or digital‑pathology pre‑screening that free clinicians for complex care.

Equip multidisciplinary teams with practical training - for example, the AI Essentials for Work bootcamp teaches promptcraft, evaluation and deployment basics for nontechnical staff - and combine that human capability with robust platforms that enforce traceability, differential privacy and logging.

A vivid rule of thumb: when a federated rollout turns a single irregular wearable ECG from a fragmented file into a validated alert, it proves the model's value in one morning rather than months of pilots.

Start small, validate locally, then scale with national alignment (ENIA), GDPR safeguards and appropriate CE‑marking or regulatory checks as projects mature.

Bootcamp	Length	Cost (early bird)	Registration
AI Essentials for Work	15 Weeks	€3,582	AI Essentials for Work – Registration