How AI Is Helping Financial Services Companies in Netherlands Cut Costs and Improve Efficiency

Across the Netherlands, banks, insurers and payment firms are already weaving AI into everyday work: fraud detection, AML screening, credit assessments and staff-facing tools that speed routine tasks, while customer-facing chatbots - ABN AMRO's “Anna” being an early example - have handled huge volumes of queries (Anna served the majority of 250,000 customers in late 2021).

Dutch regulators are watching closely: the joint AFM and DNB report on AI's impact in the financial sector stresses responsible deployment, transparency and the need to adapt supervision as models grow more advanced, and the legal landscape shifts under the EU AI Act.

A landmark Dutch court decision - pivotal for AI-driven AML - signals growing legal room for innovation in compliance; read coverage of the ruling and its AML implications in Moody's analysis of the landmark Dutch AI ruling and AML implications, but firms must still balance efficiency gains with data quality, explainability and concentration risks.

Attribute	Information
Bootcamp	AI Essentials for Work
Description	Gain practical AI skills for any workplace; learn tools, prompts, and apply AI across business functions (no technical background required).
Length	15 Weeks
Cost	$3,582 early bird; $3,942 afterwards. Paid in 18 monthly payments.
Syllabus / Register	AI Essentials for Work syllabus (15-week bootcamp) • Register for AI Essentials for Work

“…paves the way for progress that will make banking safer for everyone. It enables broader cooperation between the financial industry and online players, as effectively tackling fraud can only be done if these work together.”

Dutch financial firms are already feeling the bottom-line effects of AI: Europe-wide research shows a majority (56%) of organisations have cut costs or increased profits through AI, with an average impact of €6.24 million and more than a third reporting gains in the €5–15 million range - numbers that explain why Dutch banks and insurers are accelerating pilots and deployments.

At the same time the workforce picture in the Netherlands is mixed: 61% of Dutch workers expect AI to affect their jobs and 43% report productivity gains from AI tools, yet managers see bigger uplift than non‑executives, revealing a measurement and perception gap that can hide real ROI unless firms modernise dashboards and link AI metrics to revenue and cost KPIs.

Treating AI as a strategic, trackable investment - rather than a scatter of ad‑hoc pilots - turns experimentation into measurable savings and new revenue streams; further detail and European benchmarks are in the EY European AI Barometer 2025 report, and the Netherlands perspective is summarised in EY's local EY Netherlands AI Barometer 2025: Netherlands perspective.

It is already clear: Those who do not engage with the topic of AI will fall behind. This applies to individual employees as well as to organizations as a whole. - Adrian Ott, Partner, Forensic leader and Chief Artificial Intelligence Officer | EY Switzerland

Dutch finance teams are already seeing AI turn slow, repetitive workflows into measurable gains: 43% of respondents report higher productivity thanks to AI, with managers far more likely to notice the change than non‑managers, and 60% of Dutch firms saying AI has saved them more than €1 million (37% saved over €5 million) - findings compiled in the EY European AI Barometer 2025 report.

In practice this looks like controllers and finance teams being reshaped by the automation of financial reporting in the Netherlands, shifting effort away from number‑crunching toward higher‑value oversight, and procurement teams using contract‑clause extraction to shorten negotiation cycles.

Still, the Netherlands' broader productivity challenge - slower labour productivity growth than peer countries - means firms must pair tools with training and clear governance so efficiency gains scale across regions rather than staying inside pockets of advanced teams; notably, Dutch employees are increasingly upskilling on AI, and organisational training is on the rise, signaling a route from pilots to persistent productivity improvement.

Metric	Netherlands
Respondents reporting AI boosts productivity	43%
Managers reporting productivity uplift	56–57%
Firms saving >€1M via AI	60% (37% > €5M)

“The fact that the majority of management sees positive cost effects from the use of AI is a strong signal. AI has led to cost savings or increased revenue within companies in the Netherlands. AI pays off.” - Menno Bonninga, partner at EY in the Netherlands and AI Lead

Primary AI use cases in Dutch financial services cluster around crime-fighting, customer experience and back-office automation: regulators note AI is already used for fraud prevention, AML/CFT screening, identity verification and credit assessments in the Netherlands, while banks deploy chatbots and staff-facing assistants to speed routine work (AFM and DNB report on AI impact in the financial sector and supervision).

Real-world impact is visible - Rabobank's AI layers have become an “invisible firewall,” detecting APP and other suspicious patterns and preventing large losses (Rabobank AI fraud detection case study: the invisible firewall) - and the Dutch courts' bunq ruling has cleared legal space for AI in AML monitoring, opening the door for more automated KYC and near-real-time risk scoring (Moody's analysis of the bunq ruling and AML compliance implications).

Other practical uses include contract clause extraction for procurement and automation of financial reporting, which shift staff from repetitive processing to higher-value oversight, though institutions remain cautious about generative models and must manage data quality, explainability and concentration risks.

“…paves the way for progress that will make banking safer for everyone. It enables broader cooperation between the financial industry and online players, as effectively tackling fraud can only be done if these work together.”

Adoption in the Netherlands is rising fast but looks different depending on the lens: CBS's AI Monitor 2024 reports 22.7% of companies with ten or more workers used AI in 2024 (up nearly nine points year‑on‑year), with financial services adoption notably higher at 37.4% and text‑based tools leading the mix (text mining 13.5%, natural language generation 12.3%, speech recognition 6.5%) - see the CBS AI Monitor for details.

At the same time industry surveys paint a more expansive picture: AWS estimates roughly 180,000 Dutch companies now use AI (about 49% adoption) and even counts “one new AI implementation every four minutes,” reflecting broader sampling and definitions.

Scaling remains the challenge: many firms lack experience or skills, and pilots often stall before becoming core systems, a gap highlighted in the Gen AI benchmarking work.

The net: pockets of deep, productive use sit beside many organisations still learning how to turn AI pilots into repeatable, measurable value.

Metric	Value / Source
Companies using AI (2024)	22.7% (CBS AI Monitor 2024)
Financial services using AI	37.4% (CBS)
Top AI techs	Text mining 13.5%, NLG 12.3%, Speech recognition 6.5% (CBS)
Large firms (500+)	59.2% using AI (CBS)
AWS industry estimate	~180,000 companies; ~49% adoption; “one new AI implementation every four minutes” (AWS)

“The fact that the majority of management sees positive cost effects from the use of AI is a strong signal. AI has led to cost savings or increased revenue within companies in the Netherlands. AI pays off.” - Menno Bonninga, partner at EY in the Netherlands and AI Lead

Dutch financial firms tend to obtain AI the pragmatic way: more than half now buy and use commercially available software off the shelf, while banks and larger players commonly adapt or commission bespoke models to fit strict compliance needs and legacy systems - a pattern captured in the CBS AI Monitor 2024 (CBS AI Monitor 2024 report on AI use by Dutch companies).

In practice this looks like a mix of packaged NLP and text‑mining tools for customer service and reporting, tailored ML for risk scoring, and vendor partnerships for industry‑specific products: ING, for example, is rolling agentic AI into transaction monitoring, KYC and mortgages (with mortgages via digital agents planned for 2026), using public and behavioural data to cut routine checks and redeploy staff - a change ING says can yield roughly a 25% productivity gain and make customer due diligence “seconds” instead of days or weeks (ING agentic AI projects for transaction monitoring, KYC and mortgages (Computer Weekly)).

Fintechs like Amsterdam's Biller show the other side: cloud-native, AI-driven BNPL that runs real‑time credit and fraud checks to shrink operational burden and speed payments (Biller.AI cloud-native BNPL case study (The Payments Association)).

The result is an ecosystem where plug‑and‑play software, vendor integrations and targeted in‑house modification are all used together under tighter governance to move pilots into production.

How AI is obtained	Details / 2024
Direct commercial software	55.6% of companies obtained AI this way (CBS AI Monitor 2024)
Financial services approach	Often modify commercial software in‑house; larger firms more likely to hire external suppliers

“Customer due diligence now takes seconds versus, ‘either a day in a very positive situation, or weeks and weeks in a very negative one' … this allows staff to focus on real risk analysis rather than data gathering.” - Marnix van Stiphout, ING

Implementation headaches are real in the Netherlands: banks and insurers face skill gaps, unclear rules and thorny data quality issues that can turn promising pilots into stalled projects.

European benchmarking shows only a small share see themselves as frontrunners, while 78% acknowledge limited GenAI experience and barely a quarter have begun meaningful staff training - signals that upskilling must sit alongside tooling if productivity gains are to scale (EY European Financial Services AI Survey: AI adoption in Europe).

Regulators are not passive - Dutch supervisors are reworking oversight to keep pace with complex algorithms and societal concerns, and public sentiment is mixed (only a quarter optimistic about AI in finance, with over 60% favouring stricter supervision) as DNB and the AFM push for transparency and accountability (DNB and AFM supervisory transformation on AI regulation in the Netherlands).

At the same time the EU AI Act offers a framework to promote trustworthy adoption, but meeting its compliance, explainability and data‑protection requirements adds cost and governance burden - so firms must invest in training, ethics frameworks and careful vendor management to avoid concentration and cyber risks while protecting jobs through retraining, not abrupt displacement (PwC analysis: EU AI Act implications for the financial sector).

Metric	Value (source)
Firms acknowledging limited GenAI experience	78% (EY)
Organisations started training programmes	~25% (EY)
Have a fully functional AI ethics framework	14% (EY)
Expect up to 10% of roles could be redundant	93% (EY)

“AI is here to stay. It is happening now and will become more and more important. Then we'd better regulate it properly.” - Steven Maijoor, DNB

Regulation in the Netherlands is pragmatic: supervisors expect financial firms to use AI responsibly while reminding the sector that

regulatory objectives and standards remain independent of the technology used

- existing rules apply even where models do the work.

The joint AFM‑DNB report sets that tone, urging institutions to mind data quality, explainability and fundamental‑rights impacts and signalling that supervision will evolve as techniques do (including follow‑up symposiums and roundtables with banks, insurers and payment firms) - see the AFM and DNB joint report on AI's impact in financial services for detail.

At the same time the EU AI Act already classifies many credit and insurance systems as high‑risk, so Dutch supervisors (AFM and DNB) will police conformity while the Dutch Data Protection Authority coordinates algorithmic oversight and impact assessments; practical expectations include DPIAs, bias testing, documentation and tightened vendor governance.

Authority	Primary supervisory focus (Netherlands)
AFM	Conduct, consumer protection, transparency and market outcomes
DNB	Prudential soundness, accountability, ethics and systemic risks
Dutch DPA (including Algorithm Coordination Directorate)	GDPR compliance, algorithmic risk signalling and EU AI Act coordination

The result: a risk‑based, innovation‑friendly approach where firms must pair tech pilots with governance and proof‑points to scale safely - more on the wider legal and supervisory landscape is in the Chambers guide to AI regulation in the Netherlands.

Practical impacts in the Netherlands are already visible in day‑to‑day operations: Amsterdam's ambitious Smart Check welfare pilot ran nearly 1,600 live applications, cost roughly €500,000 and was halted after bias and performance problems emerged - an instructive example of how algorithmic pilots can bog down frontline services and erode trust (MIT Technology Review: Amsterdam Smart Check bias and performance problems); at the same time Dutch banks are hardening real‑time defences - one major bank foiled an AI‑driven deepfake audio impersonation of a senior executive - showing how AI powers both new attacks and new defences, from behavioral biometrics to adaptive risk scoring (Next IT Security: AI-driven cyber threat management and deepfake incident).

Commercial automation vendors and startups are turning pilots into productivity wins across finance: cloud‑native BNPL and document‑processing platforms speed underwriting and shrink manual checks, while fraud systems promise big drops in false positives and faster investigations (see examples of Dutch automation and fintech case studies in industry coverage).

The net operational lesson for Dutch firms is pragmatic: pair ambitious modelling with tight audits, clear thresholds and human review, or risk costly reversals that stall the very efficiency gains AI is supposed to deliver (AI automation in the Netherlands - Dutch business case studies and practical steps).

For beginners in the Netherlands the safest route is pragmatic: pick a low‑risk, high‑volume process - invoice processing, customer‑service replies or document analysis are common starting points - and run a short pilot that measures time saved, error reduction and employee acceptance; Amsterdam case studies show firms can eliminate hours of manual data interpretation and free staff for oversight.

Pair that practical approach with Dutch supervisory expectations: follow DNB's SAFEST principles (soundness, accountability, fairness, ethics, skills and transparency) and the AFM‑DNB joint guidance so pilots remain compliant with GDPR and the upcoming AI Act (Summary of Dutch AI and financial regulation - RegulationTomorrow, DNB AI guidance explained - Stibbe).

Use off‑the‑shelf or AI‑native platforms to reduce complexity, insist on DPIAs and bias testing, keep a human‑in‑the‑loop for decisions, and invest in short, role‑specific training so the whole team understands limits and controls - then scale only after clear metrics prove value (AI automation roadmap and pilot guidance - Lleverage).

This mix of small wins, documentation and regulatory alignment turns experimentation into repeatable, safe adoption.

“We take a fundamentally different approach compared to other AI platforms. Rather than focusing on the technology itself, we concentrate on the underlying challenge: enabling business experts to automate their knowledge without getting lost in technical complexity. With Lleverage, describing the problem is all it takes to begin solving it.” - Lennard Kooy, CEO, Lleverage

Conclusion - the practical path forward for Netherlands financial professionals is clear: align AI pilots with the Dutch risk‑based supervisory mindset and EU rules, build auditable governance, and invest in practical skills so efficiency gains stick.

Regulators (DNB and AFM) expect responsible deployment across soundness, accountability, fairness, ethics, skills and transparency, and the incoming EU AI Act will make robust documentation, DPIAs and bias testing non‑negotiable (see the DNB/AFM summary of AI and financial regulation in the Netherlands DNB/AFM summary of AI and financial regulation in the Netherlands).

Operationally that means pilot small, measure outcomes, keep a central model register, ensure human‑in‑the‑loop review and tighten vendor/cloud outsourcing controls, then scale only when controls, explainability and monitoring are proven; Chambers' practical checklist of governance, training and staged rollouts is a useful blueprint for this transition.

For teams short on skills, structured workplace training - for example the 15‑week AI Essentials for Work syllabus - offers role‑focused prompt and tool training to turn experimentation into measurable productivity improvements (AI Essentials for Work 15-week syllabus (Nucamp)).

Treat compliance and documentation as part of ROI: well‑governed AI is the fastest route to lasting cost savings and safer, more efficient services in the Netherlands.

Priority	Action
Governance & compliance	DNB/AFM alignment, DPIAs, EU AI Act readiness
Pilots & operations	Small pilots, human‑in‑loop, central model register, vendor/cloud controls
Skills	Role‑specific training (e.g., AI Essentials for Work syllabus)