A digital-first bank that centralises AI on one platform and a few priority domains, puts agentic AI into mortgages, and and attaches a headcount figure (about 1,250 Operations FTEs in 2026) to its digitalisation and AI push.
How is ING using AI?
ING Groep N.V., the Amsterdam-headquartered bank that reported EUR 1,160,759 million of total assets at 30 June 2026, runs AI as a centralised programme: one cloud-based data, analytics and AI platform, a central team of about 600 AI staff (around 300 in a hub in Türkiye) and a deliberately narrow set of priority domains, namely marketing, chatbots and customer contact, know-your-customer, software engineering and wholesale lending. ING says more than 90% of its generative-AI pilots reach production, contact-centre chatbots run in seven countries (with more than 80% of chats resolved without a human in four of them), and in 2026 it moved an agentic mortgage assistant from a March pilot in the Netherlands to a wider rollout, which the CEO says cuts handling time for complex cases from seven days to five. It has also put a number on the workforce effect: a decrease of about 1,250 Operations FTEs expected in 2026, among others in KYC, alongside about EUR 350 million of incremental cost efficiencies. Google is its principal AI partner.
Lead bank
ING Bank N.V.
Total assets (group, as reported, Jun 30, 2026)
€1,160,759m total assets[3]
Charter
Dutch bank holding group; supervised by the ECB under the Single Supervisory Mechanism, with De Nederlandsche Bank as national authority
Headquarters
Amsterdam, Netherlands
Flagship platform
Central AI platform on Google Cloud (VISTA data layer) and the AI Risk Committee
AI leadership
Steven van Rijswijk · Daniele Tonella · Marnix van Stiphout · Bahadir Yilmaz · Marco Li Mandri · Tom Degen · Hans Overeem
- Centralised and focused: ING decided at board level on five priority domains (hyper-personalised marketing, chatbots and straight-through processing, KYC and compliance, software engineering, wholesale lending), built on one platform; its March 2026 investor deck reports more than 100 initiatives organised around 12 themes and more than 90% of (Gen)AI pilots moved into production.
- Customer contact: GenAI chatbots are live in seven countries (the Netherlands, Belgium, Germany, Spain, Italy, Romania and Australia); more than 80% of chats are resolved without human support in four countries; the CEO puts the share of chatbot traffic handled by AI at 75%, against about 45% a couple of years earlier.
- Agentic AI in mortgages: ING's CEO says it was the first bank to roll out agentic mortgages in Europe, piloted in the Netherlands in March 2026 and scaled from June; a human still makes the final decision on every application. Complex cases fall from seven days to five, and ING plans to extend it to other markets.
- Workforce number: the March 2026 deck expects a decrease of about 1,250 Operations FTEs in 2026, among others in KYC and associated activities (about 6,000 FTEs), within roughly EUR 350 million (3%) of incremental cost efficiencies.
- Governance: a central AI Risk Committee (set up in 2025) and a Centre of Excellence oversee AI risk; ING says it vets 140 distinct risks for GenAI systems and that much of the EU AI Act was already embedded in its framework; the 2025 20-F says AI controls are being implemented to comply as high-risk obligations phase in during 2026 and 2027.
- Partners: Google is the principal AI partner; McKinsey and the hyperscalers helped build the first wave; Microsoft 365 Copilot and engineering copilots are being rolled out; Worldline and Mastercard partnered with ING on what the companies call Europe's first end-to-end agentic payment transaction (June 2026).
- Next: conversational banking, a personal assistant with agentic capabilities in the mobile app, was planned to start a gradual rollout before summer 2026 and the CEO said on 30 July it would start 'soon'; Instant Mortgages launched in Germany with a binding credit decision in under 30 minutes.
- Industry rules: in September 2026 ING co-authored, with ASB, Bank of America, Capital One, Commonwealth Bank of Australia and NatWest, a principles paper on building trust in agentic commerce (transparency, safety, privacy and data, choice, interoperability).
What is Central AI platform on Google Cloud (VISTA data layer) and the AI Risk Committee?
ING organises AI centrally: a single platform on which it builds most of its AI, all of its generative AI and its agentic AI, managed with central risk controls, guardrails and real-time monitoring. Its annual report describes a new unified data, analytics, AI and agentic environment, and CEO Steven van Rijswijk said on 23 September 2026 that the main technology partner is Google, whose VISTA data layer lets ING use its models for both processing and client-interaction use cases. About 600 people work on AI, around 300 of them in a hub in Türkiye. In 2025 ING set up a dedicated central AI Risk Committee to oversee new domains such as voicebots and agentic AI.[1][5][9]
What has ING done on AI, and when?
- Sep 23, 2026
CEO at the BofA conference: 600 AI staff, Google as main partner[5][6]
Steven van Rijswijk describes a centralised approach with clear use cases, 600 people working in AI (300 in a hub in Türkiye), Google as the main partner via VISTA, agentic mortgages cutting complex cases from seven to five days, and conversational banking and a 75% chatbot share to come.
- Sep 22, 2026
Six banks publish principles for agentic commerce[15][16]
ING co-authors a paper with ASB Bank, Bank of America, Capital One, Commonwealth Bank of Australia and NatWest calling for transparency, safety, privacy and data, choice and interoperability, including an auditable record of what a customer asked an agent to do.
- Jul 30, 2026
2Q26 results: Instant Mortgages in Germany and agentic mortgages 'live in production'[3][4]
ING launched Instant Mortgages in Germany with a binding credit decision in under 30 minutes and expanded AI-supported personalised marketing across markets; the CEO calls agentic mortgages live in production, serving more customers without adding FTEs, and says conversational banking in the mobile app will start rolling out soon.
- Jun 10, 2026
Agentic mortgage assistant begins scaled rollout[11]
After a March pilot for applications that would normally need manual assessment, ING begins rolling the assistant out across the Netherlands; it analyses applications, explains outcomes and suggests next steps while an ING employee makes the final decision.
- Jun 2, 2026
Live end-to-end agentic payment with Worldline and Mastercard[14]
An ING customer used an AI assistant to choose and buy concert tickets, with the assistant carrying out every step to payment on infrastructure based entirely in Europe; the companies describe it as Europe's first end-to-end agentic payment transaction.
- May 26, 2026
Vibe-coded trading systems reported[17]
ING's global head of electronic trading says the bank uses AI-assisted coding to build FX and credit e-trading infrastructure in hours rather than weeks, prefers Anthropic's Claude as its external model, and credits an in-house AI currency-pricing model with an initial 50% increase in large-ticket trades, as reported by Moby via Yahoo Finance.
- Mar 18, 2026
Morgan Stanley conference: 'Driving value through digitalisation, scalability and (Gen)AI'[2]
ING reports more than 90% of (Gen)AI pilots in production, more than 100 initiatives, instant consumer-credit decisions at 65% in the Netherlands, a first agentic mortgage assistant pilot live in the Netherlands, instant mortgages due in Germany within a year, and a decrease of about 1,250 Operations FTEs expected in 2026.
- Feb 26, 2026
2025 Form 20-F: AI Risk Committee and seven-country chatbots[1]
The annual report says contact centres underwent a GenAI transformation with chatbots live in seven countries, KYC models and AI-driven customer due diligence were introduced, a dedicated central AI Risk Committee was set up for voicebots and agentic AI, and data fluency training reached more than 8,700 participants; a risk factor says AI governance and model-risk controls are being implemented for the EU AI Act.
- Jul 1, 2025
COO explains the five-domain generative-AI programme[7]
Marnix van Stiphout tells McKinsey that ING chose five strategic domains at board level, went from experimentation to deployment at scale in under a year, created a faster 'highway' AI risk process led by senior colleagues, and partnered with McKinsey and hyperscalers; sensitive areas always keep humans in the loop.
Where does ING use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| Agentic mortgage assistant (Netherlands)[11][12][4][5] | An AI agent that takes over gathering and checking documents and moving cases between systems, explains outcomes and suggests next steps, with an employee deciding; piloted in March 2026 and rolling out, with other markets to follow. | Rolling out | Generative & agentic AI |
| GenAI contact-centre chatbots[1][2][9] | Chatbots live in seven countries; more than 80% of chats resolved without human support in four of them; customers can ask for a person at any point. | In production | Customer-facing chatbots |
| Conversational banking in the mobile app[2][4] | A personal assistant with agentic AI capabilities in the app, with gradual rollout starting around mid-2026. | Announced | Generative & agentic AI |
| AI-driven KYC and transaction monitoring[1][8] | A new Risk Assessment Model for customer due diligence and a secondary-analytics transaction-monitoring model live in Belgium, the Netherlands, Romania and Australia; due diligence that took a day to weeks can take seconds, according to the COO. | In production | AML / KYC |
| Instant credit decisions and AI-assisted mortgages[2][1][3] | 65% instant credit decisions for consumer loans in the Netherlands; machine-learning scorecards for mortgage credit decisions in Australia; online mortgages in Germany with a binding decision in under 30 minutes. | In production | Credit scoring & underwriting |
| Hyper-personalised marketing[2][1][3] | GenAI-powered campaigns and targeting, with more than 3% higher conversion, rolled out in six countries. | In production | Generative & agentic AI |
| GenAI in software engineering[1][9][17] | GenAI used to speed up development and improve code quality, with copilots for engineers and Microsoft 365 Copilot rolling out; AI-assisted coding is also used to build electronic-trading systems, as reported. | In production | AI-generated code & coding agents |
| In-house AI FX pricing model[17] | ING credits an in-house AI currency-pricing model with an initial 50% increase in large-ticket trades, as reported by Moby via Yahoo Finance. | In production | Trading & capital markets |
| Agentic payments with Worldline and Mastercard[14][15] | A live end-to-end agentic payment on European infrastructure, plus a six-bank principles paper that asks for audit trails and dispute processes for agent-initiated payments. | Pilot | Fraud detection |
| AI Risk Committee and Centre of Excellence[1][9] | A central AI Risk Committee for new domains such as voicebots and agentic AI, a Centre of Excellence, and vetting of 140 specific GenAI risks. | In production | AI governance (general) |
| Google as principal AI partner[5][7] | Google is the main partner through its VISTA data layer; ING also used McKinsey and hyperscalers to build its first wave of GenAI. | In production | Third-party & vendor AI |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
What has ING disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Total assets (ING Group) | EUR 1,160,759 million[3] | Jun 30, 2026 |
| Operations FTE decrease expected in 2026 | About 1,250, among others in KYC and associated activities (about 6,000 FTEs)[2] | Mar 18, 2026 |
| Incremental cost efficiencies in 2026 | -3% (about EUR 350 million) of 2026 expenses[2] | Mar 18, 2026 |
| (Gen)AI pilots moved into production | More than 90%[2] | Mar 18, 2026 |
| AI initiatives in high-value areas | More than 100, organised around 12 themes[2] | Mar 18, 2026 |
| Chats resolved without human support | More than 80% in four countries, rolled out in seven countries[2] | Mar 18, 2026 |
| Instant credit decisions for consumer loans (Netherlands) | 65%[2] | Mar 18, 2026 |
| Conversion uplift from hyper-personalisation | More than 3% higher, rolled out in six countries[2] | Mar 18, 2026 |
| Employees working in AI | About 600, including about 300 in a hub in Türkiye[5][6] | Sep 23, 2026 |
| Complex mortgage processing time (Netherlands, agentic AI) | From seven days to five[5][6] | Sep 23, 2026 |
| Chatbot traffic routed by AI rather than people | 75% with GenAI, from about 45% a couple of years earlier (CEO)[5] | Sep 23, 2026 |
| Group headcount over the past 12 months | Down by more than 1%, while income rose by more than 5% and costs by about 2%[13] | Jun 30, 2026 |
| Data fluency training participants | More than 8,700 across 18 functions in 14 countries[1] | Dec 31, 2025 |
“This isn’t about scaling a tech tool, it’s about rewiring how we work.”[7]
Marnix van Stiphout · Chief Operations Officer and Chief Transformation Officer · Jul 1, 2025
“in sensitive areas, we’ll always keep humans in the loop.”[7]
Marnix van Stiphout · Chief Operations Officer and Chief Transformation Officer · Jul 1, 2025
“Agentic Mortgages are a prime example of how our AI capabilities allow us to achieve true scalable growth.”[4]
Steven van Rijswijk · Chief Executive Officer · Jul 30, 2026
“For customers, what matters is not the technology itself, but getting clear, reliable answers they can trust”[11]
Bahadir Yilmaz · Chief Analytics Officer · Jun 10, 2026
“We have gone for a very centralized approach with clear use cases that we say, "This is what we want everybody to do so that you really can focus on the big wins."”[5]
Steven van Rijswijk · Chief Executive Officer · Sep 23, 2026
Who runs AI at ING?
Daniele Tonella
Chief Technology Officer (global head of IT since August 2024)[10]
Marnix van Stiphout
Chief Operations Officer and Chief Transformation Officer (oversees AI in operations)[7][8]
Bahadir Yilmaz
Chief Analytics Officer[11]
Marco Li Mandri
Head of Advanced Analytics Strategy[9]
Tom Degen
Head of Mortgages, ING Netherlands (agentic mortgage assistant)[11]
Hans Overeem
Head of Payments, ING Netherlands (agentic payment transaction)[14]
Which regulators govern ING's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| ECB | ING Bank N.V. is a significant institution under ECB banking supervision, whose 2026-28 priorities cover digitalisation and AI-related risk; the ECB also approves appointments to ING's Management Board. | SSM supervisory priorities 2026–28 Supervision Newsletter: AI use cases for credit scoring and fraud detection (Nov 2025) |
| EU AI Act | ING's 20-F says it is implementing AI governance and model-risk controls ahead of high-risk obligations phasing in during 2026 and 2027; credit-scoring and instant lending decisions are the high-risk category closest to its mortgage and consumer-loan automation. | Regulation (EU) 2024/1689 EBA factsheet on the AI Act GDPR Article 22 |
| EBA | EBA loan-origination and monitoring guidelines and its AI adoption work bear on automated credit decisions such as ING's instant consumer-credit decisions and AI-assisted mortgage assessment. | EBA Guidelines on loan origination and monitoring (EBA/GL/2020/06) EBA report: Rising application of AI in EU banking and payments (Sep 2025) |
| Basel Committee | ING's data strategy cites BCBS 239 alongside the EU AI Act, and its central AI platform and Google dependency engage the Basel third-party risk principles. | BCBS 239 BCBS Third-Party Risk Principles (Dec 2025) |
| DNB / AFM | DNB is ING's home prudential supervisor alongside the ECB, and the Dutch authority for model-risk and conduct oversight of automated mortgage and credit processes. |
What does the public record suggest about ING's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
Centralisation is the strategy, and the evidence
ING attributes its pilot-to-production rate (more than 90%, against an industry average its head of advanced analytics strategy puts at about 30%) to one platform, a central team and a short list of use cases, and its CEO describes the approach as 'very centralized'. The claim is self-reported, but the single-platform design is documented across the 20-F, the March 2026 deck and the CEO's remarks.
02
AI is attached to a workforce number
The March 2026 deck, titled around digitalisation, scalability and (Gen)AI, shows a decrease of about 1,250 Operations FTEs in 2026, mainly in KYC and associated activities, while the COO said in 2025 that work done with AI could need 25% fewer people with the 25% redeployed.
03
Agentic AI starts in a regulated, human-decided flow
ING chose mortgages, where an employee makes the final decision and the agent gathers, checks and explains. That keeps the first agent inside human oversight of a process the EU AI Act treats as high-risk when it involves creditworthiness, and it is the template for rolling out to other markets.
04
One principal cloud and model partner concentrates dependency
Google is named as the main AI partner and ING's head of advanced analytics strategy concedes that EU-hosted cloud still leaves some dependencies for continuity of service. That puts the platform in scope of the ECB's and Basel Committee's third-party and concentration work.
05
ING is helping write the rules for agent payments
Beyond running a live agentic payment, ING co-authored a six-bank paper asking for audit trails, authentication and dispute processes for agent-initiated payments, positioning the bank to shape liability and trust standards before regulators do.
How is ING using generative and agentic AI?
ING focuses AI on a few domains: GenAI chatbots in contact centres (live in seven countries), AI-driven KYC and customer due diligence, hyper-personalised marketing, GenAI in software engineering and support for wholesale lending. Its first agentic use is a mortgage assistant in the Netherlands that reviews applications and suggests next steps while an employee decides; a conversational, agentic assistant in the mobile app is next.
Is ING cutting jobs because of AI?
ING's March 2026 investor deck on digitalisation, scalability and (Gen)AI expects a decrease of about 1,250 Operations FTEs in 2026, among others in KYC and associated activities, which account for about 6,000 FTEs. Its COO told Computer Weekly that AI introduced into an operations process brings a 25% productivity gain and that the capacity can be redeployed to growth and more complex tasks.
Does ING have a Chief AI Officer?
The sources reviewed name no Chief AI Officer. AI is organised through the Chief Operations Officer (who oversees AI in operations), the Chief Technology Officer, the Chief Analytics Officer and a central AI team of about 600 people, with a central AI Risk Committee overseeing risk.
Which technology partners does ING use for AI?
Google is ING's principal AI partner, using its VISTA data layer. ING also worked with McKinsey and hyperscalers to build its first GenAI wave, is rolling out Microsoft 365 Copilot, and ran a live agentic payment with Worldline and Mastercard; a report in May 2026 said its electronic-trading team prefers Anthropic's Claude for coding.
- ING Groep N.V. Annual Report on Form 20-F for the fiscal year ended December 31, 2025
- Driving value through digitalisation, scalability and (Gen)AI (Morgan Stanley European Financials Conference 2026 presentation)
- ING posts 2Q2026 net result of €1,947 million, reflecting accelerated growth in customer base and customer balances
- Earnings call transcript: ING beats Q2 2026 forecasts and lifts outlook
- ING at Bank of America 31th Annual Financials CEO Conference: growth lifts targets
- ING Group Raises 2027 ROE Target Above 16% as Revenue, AI Drive Growth
- Deploying AI at speed and scale: Talking with ING’s Marnix van Stiphout
- ING Bank transforming operations through agentic AI (undated page, retrieved 2026-10-05)
- Interview: How ING reaps benefits of centralising AI (undated page, retrieved 2026-10-05)
- Interview: Daniele Tonella, CTO, ING (undated page, retrieved 2026-10-05)
- ING deploys agentic AI to speed up mortgage decisions
- ING increases use of AI in mortgage application process (undated page, retrieved 2026-10-05)
- ING Group Q2 Earnings Call Highlights
- ING completes live end-to-end European agentic payment transaction
- Building Trust in Agentic Commerce: Customer-centric principles for a trusted and thriving ecosystem
- Banks Call for Audit Trail From AI Shopping Instructions to Payment Outcome
- ING Built a Trading System in Hours With AI. Wall Street's Watching
When ING moves on AI, you'll read it here first.
when one of the banks we cover moves on AI, it is in the next morning's brief · six sourced stories · 7 am ET · free
plus every tracker, bank and agent page update, the morning after · leave any morning