Top 10 AI Prompts and Use Cases and in the Financial Services Industry in Spain

Spain's financial services sector is moving from curiosity to practical pilots: Cognizant's research finds 73% of Spanish firms are at least piloting customer‑facing generative AI, even though near‑term investment per company (~$23.5M) trails the global average and momentum scores sit ~22% below world levels; rising compute (MareNostrum 5 at 314 petaflops) and strong open‑data initiatives are helping bridge that gap.

Insurers, already heavy users of cloud and analytics, are pairing cautious governance with fast adoption - UNESPA's industry view maps AI across claims, fraud detection and underwriting while regulators and a new supervisory agency (AESIA) push for accountability and sandboxes.

For teams ready to move from pilots to production, targeted upskilling and prompt engineering are practical first steps - see Cognizant's Spain findings and a sector primer on AI in insurance for context - and short, role‑focused courses like Nucamp AI Essentials for Work bootcamp can speed readiness for regulated environments.

Metric	Spain (source)
Pilot adoption of gen AI	73% of Spanish businesses (Cognizant)
Near‑term investment per company	$23.5M vs global $47M (Cognizant)
MareNostrum 5 peak performance	314 petaflops (Cognizant)
EU open data maturity rank	7th (Cognizant)

Methodology blended three complementary strands to keep the focus firmly on Spain: a catalog-first sweep of real-world vendor use cases (the exhaustive Google Cloud roundup of generative AI deployments helped map proven patterns), practical system guidance on which models and workflows to choose (Ethan Mollick's “Using AI Right Now” guide informed model‑selection, Deep Research and multimodal options), and business‑strategy scaffolding that stresses data audits, ethics and stepwise pilots (Harvard Business School's AI strategy checklist and CIO Dive's “start small, win big” advice shaped prioritization).

Those sources were read alongside sector reporting from Cognizant to ensure recommendations match Spanish maturity levels and regulatory pressure - so the research approach favors short, measurable pilots (one clear KPI, a lean data audit, and an ethical guardrail) before scale.

The result: a reproducible, low‑risk checklist for Spanish financial teams that ties vendor capabilities to tangible pilots and training pathways rather than abstract promises; think of it as turning a sprawling vendor playbook into a three‑item sprint plan that uncovers one actionable win within 90 days.

Read the core references: Google Cloud generative AI use-case library, Ethan Mollick “Using AI Right Now” model-selection guide, and a Nucamp AI Essentials for Work syllabus on low-cost pilots in Spain.

“Over the last decade, we saw the emergence of firms that are designed and architected to release the full potential of data, algorithms, and AI.” - Harvard Business School

Automated customer service in Spain is already moving beyond simple FAQ bots into 24/7, brand‑centric experiences that cut wait times and free human agents for complex cases: Verbio's Call Automation, deployed at one of Spain's largest banks, replaced a clunky IVR, raised transcription accuracy above 90% and shrank average resolution times that once hovered around 200 seconds, enabling true self‑service and multimodal routing that saved millions and lifted CX; internal innovation programs at BBVA are now vetting ChatGPT Enterprise pilots that promise faster agent training and richer sales support, while telcos such as Telefónica are expanding AI into consumer and IoT services that can tie security, notifications and conversational channels together.

These examples show a clear Spanish pattern - combine in‑language NLU, human‑in‑the‑loop design and channel integration (WhatsApp, apps, voice) to deliver measurable agility and lower operational cost, not hypothetical gains.

“Our suite aims to empower agents in various aspects, with the goal of saving them time and ensuring they have all the tools needed to provide a unique customer experience.” - Jordan Rodríguez

Spanish banks and insurers aiming to stop fraud in its tracks are increasingly looking at GPU‑accelerated pipelines that make real‑time scoring and graph analysis practical: NVIDIA's customer stories show teams that once took weeks to stitch terabytes of transaction history can now surface suspicious networks in minutes, while IRS pilots saw up to 20x faster experiments and big cost wins; AWS's RAPIDS on Spark benchmarks similarly report up to ~10.5x speedups and dramatic infrastructure savings, enabling rolling‑window features and GNN embeddings to run fast enough for live transaction blocking without adding customer friction.

Firms such as HSBC are already using AI and ML to catch money‑laundering rings that legacy rule engines miss, and Spanish players can follow that playbook by pairing RAPIDS‑style ETL, graph neural nets and low‑latency GPU inference to reduce false positives and keep friction low for honest customers.

For teams planning pilots, the practical payoff is clear: what used to be post‑mortem forensics becomes near‑instant prevention, turning a flood of alerts into a focused queue of high‑confidence investigations (NVIDIA fraud detection customer stories and case studies, AWS RAPIDS Accelerator for Apache Spark fraud detection blog).

Source	Reported Speedup	Reported Infra Cost Reduction
NVIDIA / IRS case	20×	50%
AWS RAPIDS benchmarks	~10.5×	up to 88%

“The Cloudera and NVIDIA integration will empower us to use data-driven insights to power mission-critical use cases. We're currently implementing this integration, and are already seeing over 20x speed improvements at half the cost for our data engineering and data science workflows.” - Joe Asaldi

Credit risk assessment and underwriting are getting a practical boost from alternative data, and Zest AI is a clear example of how lenders can move past thin bureau files: its platform ingests cash‑flow signals such as utility and rent payments and mobile operator fees, and analyzes features like delinquency duration and gaps between payments to sharpen risk prediction - think of those recurring bills as the missing puzzle pieces that turn an anonymous application into a readable repayment profile.

For Spanish banks and fintechs wrestling with under‑banked or “thin‑file” customers, that means better inclusion without loosening standards: expanded data can cut unscorable populations and lift approval rates (Equifax reports up to a 60% reduction in unscorables and 20%+ more approvals when alternative data is used).

Zest AI is sold via custom pricing with demos rather than free trials, so a short PoC that validates local data sources (rent, utilities, telco records) against portfolio performance is a practical first step - see the Zest AI alternative data provider profile (RiskSeal) and Equifax's guidance on using alternative credit data to build fairer, more predictive underwriting pipelines.

Provider / Source	Key alternative data	Pricing / Trial
Zest AI alternative data provider profile (RiskSeal)	Cash flow: utility & rent payments, mobile operator fees; delinquency duration, payment gaps	Custom pricing; demo available; no free trial
Equifax alternative data guidance for credit risk evaluation	Telco & utility payment insights, bank transactions, income data - expands scoring and inclusion	Products & consultations available

Algorithmic trading and portfolio management in Spain are steadily adopting whole‑portfolio tech rather than point solutions: BlackRock's Aladdin brings unified risk models, portfolio construction and stress‑testing into the same workflow that firms use to manage trillions globally, and European tools like BlackRock Aladdin platform and Portfolio 360 powered by BlackRock Aladdin expose those capabilities to local asset managers and insurers so they can run consistent asset‑allocation, ESG screening and scenario analysis across public and private books; Spanish adopters from CaixaBank's Aladdin Wealth advisory to MAPFRE AM's platform transformation show the local appetite for integrated analytics that cut operational complexity and improve regulatory reporting readiness.

The practical payoff for Spanish teams is straightforward: clearer risk‑drivers, repeatable portfolio reviews and faster, defendable recommendations for clients and boards - turning fragmented data into a single source of truth that supports compliance (Solvency II, SFDR) and scaled decisioning without chasing siloed spreadsheets.

For firms thinking about production pilots, Aladdin's mix of analytics, optimisation and client tools offers a way to industrialise portfolio‑level alpha and governance in one stack (Portfolio 360 powered by BlackRock Aladdin).

Metric	Value / Example
EMEA portfolios analysed (Portfolio Consulting)	6,000+ portfolios
EMEA AUM consulted	USD 700BN
Spanish adopter (press)	MAPFRE picks Aladdin (Dec 2, 2024)

“One of the founding philosophies of Aladdin has been more eyes on the data leads to better outcomes for the enterprise.” - Mark Paltrowitz

For Spanish firms such as Telefónica and Iberdrola, personalization is no longer just targeted ads - it's a way to fuse customer value with compliance and resilience by turning every interaction into a timely, auditable action; platforms described by Latinia show how hyper‑personalized alerts can satisfy PSD3/DORA traceability while routing critical messages across channels so a notification becomes both customer help and documentary evidence for auditors (Latinia: hyper-personalization for banking regulatory compliance).

Backing this with behavioral science - building dynamic segments from transaction, geolocation and interaction data and nudging customers with context-aware offers or fraud warnings - boosts relevance and trust, not just click rates; practical guides argue that combining AI, real‑time data and tested nudges lets banks and utilities move from generic blasts to proactive, measurable journeys (Data Management Blog: behavioral science and data-driven hyper-personalization in banking).

The memorable payoff: a single, well‑timed message can be the difference between a delighted, retained customer and a regulatory headache - so aim for precision, traceability, and channels the customer actually prefers.

“Consumers often make financial decisions based on behavioral biases rather than pure rationality. Understanding the psychological factors as to why decisions are made, such as loss aversion or herd mentality, can enhance the effectiveness of teams in designing customer-centric solutions,” says Gartner.

Regulatory compliance in Spain is no theoretical exercise - it's a high‑stakes operational reality where the EU's 2024 AML package, a newly expected EU authority (AMLA) and active national supervisors drive tougher inspections and larger fines, so teams must turn policy into auditable, machine‑ready controls; Uría Menéndez's primer explains the legal backbone and the Commission/SEPBLAC supervisory role that makes enforcement intense, while practical checklists (KYC, enhanced due diligence, transaction monitoring, appointed SEPBLAC delegates and internal control bodies) form the day‑to‑day playbook.

That's where surveillance tech pays for itself: natural‑language tricks and fuzzy matching now support sanctions screening and rapid dossier assembly (Citco's pilots used NLP to flag birthplace strings like

Moscow

from document repositories), and advances in Spanish NER models make extracting names, addresses and entity roles from unstructured filings far more reliable - see work on Spanish NER for semantic annotation and a hands‑on look at NLP for sanction screening to plan realistic pilots.

The bottom line: embed traceability and low‑latency alerts into existing AML controls, run short PoCs that map NLP outputs to investigator workflows, and treat every automated hit as evidence to be defended in the next audit.

Topic	Key points / sources
Legal framework	EU AML package 2024 and Spanish anti‑money laundering rules - Lexology analysis
Supervisors & enforcement	Commission for the Prevention of Money Laundering; SEPBLAC - high inspection and sanction activity
Tech for surveillance	NLP for financial sanctions screening and innovation - AIMA article and Spanish NER models for entity extraction

Back‑office teams in Spain can stop treating paper and PDFs as a cost center: modern Document AI - from layout analysis and OCR to named‑entity extraction and LLM‑based summarization - makes invoices, claims and KYC packs machine‑readable so workflows move from days to hours.

Dataiku's practical playbook describes how combining layout analysis, OCR (or multimodal M‑LLMs), and structured extraction yields reliable fields for downstream automation, while RPA + OCR orchestration handles routing, validation and system updates without manual copy‑paste.

That mix is especially valuable in Spain's multilingual, regulation‑heavy environment: native language detection, translation and Spanish NER pipelines let teams extract names, dates and contract clauses with audit trails suitable for compliance.

Think of a three‑inch stack of claims turning into a searchable database by lunchtime - freeing specialists for exception handling instead of data entry. For hands‑on guidance, see Dataiku NLP documentation for natural language processing and the Dataiku Document AI overview for document understanding, and pair that with low‑cost pilot playbooks to prove ROI quickly.

Financial forecasting and predictive analytics are the bridge between Spain's macro outlook and day‑to‑day risk decisions: the EU Spring 2025 forecast pegs real GDP growth at 2.6% in 2025 (softening to 2.0% in 2026), with headline inflation easing to 2.3% in 2025 and unemployment drifting below 10% by 2026 - numbers that underwrite loan‑loss scenarios, liquidity stress tests and revenue projections (EU Spring 2025 macro forecast for Spain).

Those medium‑term signals matter, but so does higher‑frequency data: CaixaBank's real‑time economics work highlights how daily consumption and payment flows can fine‑tune nowcasts and early‑warning models (CaixaBank Research real‑time indicators), and the rapid growth of Spain's instant‑payments ecosystem means predictive pipelines must ingest streaming payments to catch inflection points (Spain real‑time payments market outlook).

The practical takeaway: combine the EU's scenario baselines with live transaction signals and short‑horizon models so dashboards flag trouble before quarterly reports arrive - turning a 2.6% growth forecast from a calendar note into real operational decisions backed by data.

Indicator	2024	2025	2026
GDP growth (%, yoy)	3.2	2.6	2.0
Inflation (%, yoy)	2.9	2.3	1.9
Unemployment (%)	11.4	10.4	9.9
General government balance (% of GDP)	-3.2	-2.8	-2.5
Gross public debt (% of GDP)	101.8	100.9	100.8
Current account balance (% of GDP)	3.1	2.7	2.8

MAPFRE has pushed computer vision, drones and document AI from pilots into everyday claims and underwriting workflows in Spain - using mobile‑photo appraisals with Tractable to speed auto repairs (cutting turnaround by up to two weeks) and drone photogrammetry with PwC Spain to create 3D site models accurate to ~1cm for faster, indisputable loss assessment; at the same time MIA GPT and virtual agents have put thousands of documents and routine queries at employees' fingertips, with more than 3,000 users, 1,500 consultable documents and satisfaction north of 80%, while data‑driven process redesign can compress simple claim handling to seconds rather than days.

These advances are paired with MAPFRE's humanistic AI manifesto and internal risk tooling to catch bias and protect privacy, so automation frees experts for exceptions instead of replacing them.

For Spanish insurers the lesson is practical: combine vision, aerial imagery and NLP for richer evidence, run tight PoCs against local data, and bake governance into every deployment (MAPFRE on AI and responsible deployment, MAPFRE + Tractable auto‑claim appraisals, MAPFRE drone PoC with PwC Spain).

Metric	Value / Source
Spanish customers interacting with AI	70% of MAPFRE's 7.7M customers (MAPFRE)
MIA GPT adoption	~3,000 users; 1,500 documents; >80% satisfaction (MAPFRE)
Auto‑claim operational impact	Repair times cut by up to two weeks with Tractable
Drone imaging accuracy	3D models with ~1 cm resolution (PwC/MAPFRE PoC)

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

Spain's cyber posture for financial services is sharpening around behavior‑centric detection: INCIBE's deep dive on UEBA shows how profiling users, devices and services with ML turns mountains of logs into actionable signals - real‑time anomaly scoring spots oddities (think “finding the one fish that swims against the school”) that rule‑based systems miss, lowers false positives, and gives SOCs a shorter, higher‑confidence queue for investigation (INCIBE report: behavior analysis for cybersecurity and UEBA in Spain).

Practical Spanish deployments pair UEBA with SIEM and SOAR to trace lateral movement, privileged‑account misuse and rare exfiltration patterns, while dynamic peer‑group baselines and temporal models reduce alert fatigue.

INCIBE also stresses GDPR guardrails - data minimization, transparency and EEA residency or safeguards for transfers - so detection gains don't come at the cost of privacy compliance.

For banks and insurers juggling instant payments and large customer datasets, integrating UEBA into existing controls is a fast, defendable way to turn suspicious noise into prioritized, auditable leads (Securonix blog: behavioral analytics telltale signs of attack).

Spain's path to practical AI in financial services is straightforward if treated like a check‑list rather than a leap: run a tight pilot with one clear KPI and aim to surface “one actionable win within 90 days” (start small, measure fast - Cognizant's Spain study shows 73% are already piloting but investment and momentum lag), pair that pilot with role‑focused upskilling so teams can own prompts and workflows (short courses such as the Nucamp Nucamp AI Essentials for Work bootcamp teach usable prompt craft and business‑facing AI skills), bake compliance and auditability into design from day one to align with AESIA, the RD Sandbox and Spain's draft national AI law, and lock down data hygiene and compute choices so models run on trusted, GDPR‑compliant pipelines (Spain's compute and open‑data strengths are a clear accelerator).

Close the loop by measuring revenue, quality and time‑savings (Cognizant/EY metrics show productivity and revenue lift justify projects), partner where talent gaps exist, and document every step so regulators and auditors can trace decisions - a practical, repeatable checklist that turns cautious pilots into defensible production services without overpaying for hype.

Checklist item	Quick action & source
Tight pilot with one KPI	Run 90‑day PoC; measure a single metric (adoption/throughput) - Cognizant study
Upskill staff	Role‑focused training on prompts and tools - Nucamp AI Essentials for Work bootcamp
Governance & compliance	Design for AESIA, RD Sandbox & Draft Spanish AI Law from day one - JDSupra AI Watch
Data & compute hygiene	Use GDPR‑aligned pipelines and Spain's open data/compute assets to reduce risk - Cognizant