A cloud-first transformer turning AI into a cost lever: after moving its core IT to Google Cloud and industrialising 153 AI use cases in 2022-2025, Italy's largest bank is betting its 2026-2029 plan on agentic AI to lift back-office productivity by more than 20% while about 12,400 staff leave by 2029.
How is Intesa Sanpaolo using AI?
Intesa Sanpaolo, Italy's largest banking group with EUR 989,207 million of total assets at 30 September 2026, treats AI as the second stage of a cloud migration. Its 2026-2029 Business Plan, presented on 2 February 2026, lists 'AI/GenAI and Agentic AI evolution' among its cost levers and targets more than 20% higher middle and back-office productivity and automated resolution of more than 80% of Digital Branch client requests by 2029, up from about 30% in 2025. The AIxeleration programme industrialised 153 AI use cases in 2022-2025 with about 300 dedicated AI specialists, and 33 AI, GenAI and agentic AI projects had been launched under the new plan by 30 September 2026. The base is isytech, a cloud-native platform: the core IT migration to Google Cloud and TIM finished in July 2026, about 71% of applications were cloud-based in September, and the bank says it has signed hyperscaler agreements for cloud and AI services for 2026-2032. Massimo Proverbio, Chief Data, AI and Technology Officer, runs the programme, which the plan pairs with about 12,400 staff exits and 6,300 young hires by 2029. In 2026 Italy's data protection authority fined the bank EUR 17.6 million over the customer profiling behind its Isybank migration, and its Fideuram unit lost EUR 95 million to an AI voice-cloning scam, according to Reuters sources.
Lead bank
Intesa Sanpaolo S.p.A.
Total assets (group, as reported, Sep 30, 2026)
EUR 989,207 million total assets[2]
Charter
Italian banking group; Intesa Sanpaolo S.p.A. is a significant institution directly supervised by the European Central Bank under the Single Supervisory Mechanism, with the Bank of Italy as national competent authority. Group companies include Isybank (digital bank), Fideuram (private banking) and the banks of the International Banks Division in Central and Eastern Europe
Headquarters
Turin and Milan, Italy
Flagship platform
isytech (cloud-native platform on Google Cloud) and the AIxeleration programme
AI leadership
Carlo Messina · Massimo Proverbio · Gian Maria Gros-Pietro · Marco Ditta · Domenico Fileppo · Matteo Feraboli · Valerio Cencig · Stefano Lucchini · Walter Chiaradonna · Marcello Mentini · Piero Boccassino · Mario Costantini · Antonio Valitutti
- Plan targets: the 2026-2029 Business Plan (2 February 2026) names AI, GenAI and agentic AI as one of five cost levers, with middle and back-office productivity up more than 20% and automated resolution of Digital Branch client requests above 80% in 2029, from about 30% in 2025 (37% at 30 September 2026).
- Use-case record: AIxeleration, launched in 2023, industrialised 153 AI use cases by the end of 2025 (49 of them in 2025, against a plan target of 150) with about 300 dedicated AI specialists; 33 AI, GenAI and agentic AI projects were launched in the first nine months of 2026.
- Cloud first: isytech, the internally developed cloud-native platform behind the Isybank digital bank, is being extended to the whole group; cloud-based applications rose from about 10% at end-2021 to 64% at end-2025 and about 71% at 30 September 2026, with about 100% targeted by 2029.
- Spending: EUR 5.6 billion of technology investment and about 2,430 IT specialists hired in 2022-2025, plus EUR 4.6 billion of technology and growth investment planned for 2026-2029; Proverbio puts the two-cycle total above EUR 10 billion.
- Workforce: about 12,400 exits and 6,300 young hires by 2029, a net cut of about 6,100, with savings of about EUR 570 million at run rate in 2030 that the bank attributes to technology and AI-enabled process streamlining; a union-bank Digital Transformation, AI and Privacy Committee monitors the effects.
- Digital Branch: about 18 million contacts in 2025, 4 to 5 million handled by a virtual agent with accuracy above 90%, and every human contact supported by a GenAI copilot; the bank reports generative and agentic AI applications already live there in 2026.
- Governance: Group Guidelines for the Use of AI (effective April 2025) bar prohibited AI practices, require human oversight of every high-risk system and guardrails for generative AI; the annual report says the group ranked third in the world for responsible-AI publications in the 2025 Evident AI Index.
- Risk events in 2026: Italy's data protection authority fined the bank EUR 17.6 million on 12 March over the profiling of about 2.4 million customers moved to Isybank, and Fideuram lost EUR 95 million (about EUR 53 million recovered) after fraudsters used AI to imitate a lawyer's voice; the Q3 2026 deck also records an action plan against threats from frontier AI models.
- Partners on record: Google Cloud and TIM (cloud and Gemini through Vertex AI in risk management), Microsoft (Sentinel and Copilot for Security), Thought Machine (Isybank core), iGenius, Aptus.AI, Swift (federated-learning fraud trial) and Fujitsu's Frontria consortium; the hyperscalers in the 2026-2032 agreements are not named.
- Benchmark: Evident's published AI Index table for 2026 ranks Intesa Sanpaolo 27th of the 50 banks it scores (retrieved 19 November 2026).
What is isytech (cloud-native platform on Google Cloud) and the AIxeleration programme?
isytech is the bank's internally developed cloud-native core platform, deployed first through the digital bank Isybank (launched in June 2023 on Thought Machine's Vault Core in Google Cloud's two Italian regions) and now being extended across the group. On 2 July 2026 it said it had completed the migration of its core IT systems to the Google Cloud regions in Turin and Milan hosted in TIM's data centres, moving more than 800 applications, and about 71% of applications were cloud-based at 30 September 2026 against a target of about 100% by 2029. AI runs through AIxeleration, launched in 2023, which industrialised 153 use cases with about 300 dedicated AI specialists by the end of 2025 on a first MLOps platform and a group data platform on Google BigQuery. The 2026-2029 Business Plan adds an 'Agent-first' redesign of operational processes, AI Digital Expert Agents for relationship managers and agentic AI for KYC, AML and software development; 33 AI, GenAI and agentic AI projects had been launched by 30 September 2026, and the bank says it has finalised strategic agreements with global hyperscalers for cloud and AI services for 2026-2032.[7][28][37][1][4][5][2][3]
Which technology partners does Intesa Sanpaolo work with on AI?
Google Cloud is the partner with the most disclosed AI and cloud work at Intesa Sanpaolo: with TIM it hosts the bank's core IT in Italian cloud regions, and its Vertex AI and Gemini run in the risk team's data lab. Microsoft supplies Sentinel and Copilot for Security, Thought Machine the Isybank core, and iGenius and Aptus.AI generative and regulatory tools. The bank also joined Swift's federated-learning fraud trial and Fujitsu's Frontria consortium, and its venture arm invested in Straiker. The 2026-2032 hyperscaler agreements are not named, and the sources reviewed for this page did not include a patent search.[7][34][35][37][15][16][32][12][9][3]
| Partner | What it is | Status | Date |
|---|---|---|---|
| Google Cloud and TIM | Core IT migration of more than 800 applications to Google Cloud regions in TIM's Turin and Milan data centres; Democratic Data Lab on Vertex AI, BigQuery and Gemini | In production | 2026-07-02[7][34][30] |
| Microsoft | Sentinel SIEM and Copilot for Security in the Global Cyber Defense Center, according to Microsoft | In production | Early 2024[35] |
| Thought Machine | Vault Core behind Isybank; the bank also invested in Thought Machine's Series D | In production | 2023-06-15[37][28] |
| iGenius | Crystal generative business-intelligence tool, tested at Fideuram | Announced | 2023-09-29[15] |
| Aptus.AI | Machine-readable regulation for the group regulatory agenda, with Deloitte Risk Advisory | In production | 2023-10-09[16] |
| Swift | Federated-learning fraud-detection trial with 13 banks | Pilot | 2025-09-15[32] |
| Fujitsu (Frontria) | Consortium on generative-AI risks and disinformation | Member | 2025-12-02[12] |
| Straiker | Agentic-AI security start-up; Neva SGR joined its USD 64 million Series A | Invested | 2026-08-05[9] |
What has Intesa Sanpaolo done on AI, and when?
- Nov 3, 2026
9M 2026: 33 AI projects launched, 71% of applications in the cloud[3][2][19][33]
The results deck lists 33 AI, GenAI and agentic AI projects, hyperscaler agreements for cloud and AI services over 2026-2032, Digital Branch automation at 37%, an action plan against threats from frontier AI models, a machine-learning early-warning prototype for credit risk and the AI-based Connecting Markets platform launched in late July. The CEO's statement says the ECB and other authorities approved the Monte dei Paschi acquisition and that a technology hub is planned in Siena; the results call transcript contains no discussion of AI.
- Sep 28, 2026
Fideuram loses EUR 95 million to AI voice scam[25]
Private Banker International, citing Reuters sources, reported that fraudsters imitating the CEO on WhatsApp and using AI to reproduce a lawyer's voice led Fideuram's then chairman to order transfers abroad; about EUR 53 million was recovered and Intesa Sanpaolo and Fideuram declined to comment.
- Aug 5, 2026
Neva SGR invests in agentic-AI security start-up Straiker[9]
The group's venture capital company joined Straiker's USD 64 million Series A through its Neva II Fund; Straiker builds tools to discover, test and protect AI agents in production.
- Jul 2, 2026
Core IT migration to Google Cloud completed[7][31]
More than 800 applications were moved to Google Cloud's Turin and Milan regions in TIM's data centres and as many decommissioned, with more than 3,000 employees trained and more than 170 Google Cloud certifications; the bank described the infrastructure as AI-ready.
- Apr 21, 2026
AI algorithm flags potential tax fraud in public-private pilot[8]
With the Anti-Financial Crime Digital Hub, the Guardia di Finanza, the Anti-Mafia Investigation Directorate, the Bank of Italy and the Financial Intelligence Unit, the bank tested an AI algorithm on more than 40,000 corporate customers in Piedmont and Valle d'Aosta and identified more than 30 cases of potential tax fraud.
- Mar 30, 2026
2025 annual report: 153 use cases, AI compliance framework[1]
The report records 153 AI use cases industrialised since 2022, about 300 AI specialists, an AI Academy and Data and AI Community of more than 9,000 colleagues, 42,905 enrolments in the AI and Data faculty, the first MLOps platform, an AI compliance framework under the AI Act and AI risk metrics in the group's risk appetite.
- Mar 27, 2026
SofIA market-risk agent wins ABI innovation award[11]
SofIA (Self-Organized Financial Intelligent Agent), built with AI, machine learning and cloud computing under the Digital VaR programme, cut daily calculation times by 80% and functional analysis times by 70%.
- Mar 12, 2026
EUR 17.6 million privacy fine over Isybank profiling[24]
Italy's data protection authority fined the bank for unlawfully processing data of about 2.4 million customers moved to Isybank, which it had profiled by age under 65, frequency of digital-channel use and investment holdings; the authority took into account the bank's non-intentional conduct.
- Feb 26, 2026
Proverbio: more than EUR 10 billion in technology by 2029, about 150 AI use cases in production[6]
In an interview republished by the bank, the Chief Data, AI, Innovation and Technology Officer said AI will move into anti-money laundering, credit, controls and software development, that about 150 AI use cases are in production, and that management productivity rose 23% during the transformation.
- Feb 2, 2026
2026-2029 Business Plan makes agentic AI a cost lever[4][5][23][22]
The plan targets more than 20% higher middle and back-office productivity, Digital Branch automated resolution above 80% by 2029, EUR 4.6 billion of technology and growth investment and about 12,400 exits with 6,300 hires; Dow Jones put the net cut at about 6,100 jobs, and Reuters reported the cost-income ratio target of 36.8% in 2029 from 42.2%.
- Dec 2, 2025
Joins Fujitsu's Frontria consortium on AI disinformation[12]
Intesa Sanpaolo joined more than 50 organisations in Frontria, a consortium launched by Fujitsu to address risks from generative AI, including disinformation, and to align work on AI trustworthiness with the EU AI Act.
- Sep 15, 2025
Swift federated-learning fraud trial[32]
Intesa Sanpaolo was among 13 banks, with ANZ and BNY, in Swift trials that combined privacy-enhancing technologies and federated learning; trained on ten million synthetic transactions, the shared model found known frauds twice as effectively as a single-institution model.
- Mar 11, 2025
AIxeleration counts 190 use cases developed[13]
A bank article said AI has been built into processes since 2019 along four lines (data governance, data strategy, responsible AI and an AI Delivery Center) and that AIxeleration had developed 190 use cases, more than half of them in implementation.
- Oct 23, 2024
9,000 exits agreed 'in the digitalisation and artificial intelligence scenario'[17][27][26]
The bank agreed with unions on 4,000 voluntary exits in Italy within 9,000 group exits by 2027 and 3,500 young hires by June 2028, expecting about EUR 500 million a year of staff-cost savings from 2028; Bloomberg framed it as an AI push, and Reuters reported a new committee to monitor the impact of digitisation.
- Mar 29, 2024
Group data and AI officer expects first GenAI in production by end-2024[14]
Marco Ditta said generative-AI experiments had produced 15 potential use cases, that AI tools were helping global relationship managers raise proposal conversion by 70%, and that the customer-service chatbot is aided but not wholly managed by AI.
- Oct 9, 2023
Aptus.AI for regulatory analysis[16]
With Deloitte Risk Advisory and the Italian start-up Aptus.AI, the Group Regulatory Evolution Agenda Department moved to a machine-readable operating model for tracking new regulation, which the bank called a first for a European bank.
- Sep 29, 2023
iGenius chosen for generative business intelligence[15]
The bank said it would adopt iGenius's Crystal, a GPT-based generative business-intelligence tool already tested at Fideuram Private Banking, to query large amounts of data in natural language.
- Jun 15, 2023
Isybank launches on Thought Machine and Google Cloud[28][37][21]
Intesa launched Isybank, a digital bank running on Thought Machine's Vault Core in Google Cloud's Italian regions, 12 months after signing with Thought Machine; Reuters reported the bank forecast savings of about EUR 800 million a year in 2026-2027 and said it would try to move its entire business onto the cloud if Isybank succeeded.
- May 16, 2023
Lisa reads supervisory publications with machine learning[29]
The bank introduced Lisa (Linguistic Intelligence for Supervisory Awareness), a machine-learning tool that scans thousands of banking-supervision publications for correlations and trends, with a dedicated team validating its results.
- May 21, 2020
Memorandum with TIM and Google Cloud, including an AI centre[30]
Intesa Sanpaolo signed a memorandum of understanding with TIM and Google Cloud to accelerate the bank's digitalisation; the project included opening a centre dedicated to artificial intelligence, training and support for start-ups.
Where does Intesa Sanpaolo use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| AIxeleration programme[1][6] | Group programme launched in 2023 to scale AI across divisions: 153 use cases industrialised by end-2025 in areas such as HR support, the Digital Branch, regulatory analysis, technical support and coding, about 300 dedicated AI specialists and a first MLOps platform for generative-AI use cases. | In production | Generative & agentic AI |
| 'Agent-first' redesign of operations[5][3] | Plan initiatives include a conversational agentic interface for colleagues, AI Digital Expert Agents for relationship managers, agentic AI for rating assignment and validation and for drafting contracts and collateral documents, and AI tools for HR and the help desk; 33 projects were launched by 30 September 2026. | Rolling out | Generative & agentic AI |
| Digital Branch virtual agent and GenAI copilot[5][3] | Of about 18 million contacts in 2025, 4 to 5 million were handled by a virtual agent with accuracy above 90%, and all human contacts were supported by a GenAI copilot; automated resolution of client requests rose from about 30% to 37% by September 2026, with more than 80% targeted for 2029. | In production | Customer-facing chatbots |
| Agentic AI across the software development lifecycle[5][3] | The plan uses agentic AI for requirements definition, software development and testing; the Q3 2026 deck says the bank is using frontier AI capabilities to accelerate the modernisation of its software development lifecycle. | Rolling out | AI-generated code & coding agents |
| Democratic Data Lab for risk management on Google Cloud[34] | Built from 2021 on Vertex AI, BigQuery and Looker; according to Google Cloud the lab cut the time to complete EU-wide stress-test reports by more than 80% and the time to extract and load risk-model data by 30% with Gemini 1.5. | In production | Model risk management |
| SofIA market-risk agent[11] | Predictive models using AI, machine learning and cloud computing for investment-portfolio risk analysis under the Digital VaR programme; daily calculation times fell 80% and functional analysis times 70%. | In production | Trading & capital markets |
| Machine-learning early warning and agentic rating support[2][5] | A first enhanced prototype of early-warning models for credit risk using machine learning and market data was developed in 2026; the plan adds GenAI for proactive portfolio management and agentic AI for rating assignment. | Pilot | Credit scoring & underwriting |
| Anti-Financial Crime Digital Hub algorithms[8][1][20][5] | The consortium formed in 2022 with the Innovation Center and Turin's two universities developed an AI algorithm, tested on more than 40,000 corporate customers, that found more than 30 potential tax-fraud cases; the plan adds agentic AI Digital Experts for KYC and AML. | Pilot | AML / KYC |
| Real-time AI fraud detection and Swift federated learning[1][32] | The annual report describes a real-time transaction-analysis platform based on AI and machine learning and the use of the EBA Clearing FPAD risk score for corporate transfers; the bank also took part in Swift's federated-learning fraud trial. | In production | Fraud detection |
| Microsoft Sentinel and Copilot for Security[35][3] | The Global Cyber Defense Center replaced its on-premises SIEM with Microsoft Sentinel and joined the Copilot for Security early access programme in early 2024, using it to generate KQL queries and executive briefings, according to Microsoft; a 2026 action plan responds to threats from frontier AI models. | In production | Cybersecurity |
| Lisa and Aptus.AI for supervisory and regulatory analysis[29][16] | Lisa reads thousands of banking-supervision publications with machine learning (2023); Aptus.AI turns regulation into machine-readable documents for the group's regulatory agenda (2023). | In production | AI governance (general) |
| iGenius Crystal generative business intelligence[15] | GPT-based natural-language analytics tested at Fideuram Private Banking and chosen for group business intelligence in September 2023; no later rollout figure was found. | Announced | Generative & agentic AI |
| Group Guidelines for the Use of AI and AI compliance framework[10][1] | Guidelines effective April 2025 bar prohibited AI practices, apply the strictest requirements to high-risk systems, require human-in-the-loop or human-on-the-loop oversight with the business owner accountable, and require guardrails for generative AI; the board approves the guidelines and AI risk tolerance. | In production | AI governance (general) |
| Google Cloud and TIM core migration; 2026-2032 hyperscaler agreements[7][3] | More than 800 applications moved to Google Cloud regions in TIM's Italian data centres (July 2026); the bank says it finalised strategic agreements with global hyperscalers for cloud and AI services over 2026-2032 without naming them. | In production | Third-party & vendor AI |
| Customer profiling and personal-data controls[24][3] | The data protection authority's EUR 17.6 million fine concerned profiling used to move about 2.4 million customers to Isybank; the bank's NEMO programme to strengthen the control and protection of customers' personal data is at an advanced stage. | Rolling out | Data & privacy |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
What has Intesa Sanpaolo disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Total assets | EUR 989,207 million[2] | Sep 30, 2026 |
| AI use cases industrialised under AIxeleration | 153 since 2022, of which 49 in 2025 (plan target 150)[1] | Dec 31, 2025 |
| Dedicated AI specialists | About 300[1] | Dec 31, 2025 |
| AI, GenAI and agentic AI projects launched under the 2026-2029 plan | 33[2][3] | Sep 30, 2026 |
| Middle and back-office productivity target | More than 20% higher in 2029 than in 2025[4][5] | Feb 2, 2026 |
| Digital Branch automated resolution of client requests | About 30% (end-2025), 37% (September 2026), more than 80% targeted for 2029[4][3] | Sep 30, 2026 |
| Digital Branch contacts handled by the virtual agent | 4 to 5 million of about 18 million, accuracy above 90%[5] | Dec 31, 2025 |
| Cloud-based applications | About 10% (end-2021), 64% (end-2025), about 71% (September 2026), about 100% targeted for 2029[4][6][2] | Sep 30, 2026 |
| Technology investment | EUR 5.6 billion in 2022-2025; EUR 4.6 billion of technology and growth investment planned for 2026-2029[4][1] | Feb 2, 2026 |
| IT specialists hired | About 2,430 in 2022-2025[1][5] | Dec 31, 2025 |
| Planned workforce change | About 12,400 exits and 6,300 young hires by 2029 (net about 6,100); about EUR 570 million savings at run rate in 2030[4][23] | Feb 2, 2026 |
| AI and Data faculty enrolments | 42,905 participants[1] | Dec 31, 2025 |
| Applications migrated to Google Cloud | More than 800, with more than 3,000 employees trained and more than 170 certifications[7][31] | Jul 2, 2026 |
| SofIA market-risk agent | Daily calculation times down 80%, functional analysis times down 70%[11] | Mar 27, 2026 |
| AI tax-fraud pilot | More than 40,000 corporate customers tested; more than 30 potential tax-fraud cases identified[8] | Apr 21, 2026 |
| Isybank customers | About 1.3 million[2] | Sep 30, 2026 |
| Privacy fine over Isybank profiling | EUR 17.6 million; about 2.4 million customers affected[24] | Mar 12, 2026 |
| Fideuram AI voice scam | EUR 95 million transferred, about EUR 53 million recovered[25] | Sep 28, 2026 |
“Technological innovation and artificial intelligence will play a decisive role in creating new opportunities for future generations.”[1]
Gian Maria Gros-Pietro · Chair of the Board of Directors, 2025 annual report · Mar 30, 2026
“By partnering with Google Cloud and TIM, we changed technology, reduced costs and, at the same time, laid the foundations for building Isytech, the cloud-native digital technology platform serving customers and colleagues across our Group.”[7]
Massimo Proverbio · Chief Data, AI and Technology Officer · Jul 2, 2026
“To my knowledge we are the only large European bank with a business plan delivering cost reduction,”[22]
Carlo Messina · Managing Director and Chief Executive Officer, to analysts on the 2026-2029 plan · Feb 2, 2026
“We have a lot of responsibility using AI,”[14]
Marco Ditta · Executive director and group data and artificial intelligence officer · Mar 29, 2024
“I think that by the end of 2024, we will see the first generative AI solution in production.”[14]
Marco Ditta · Executive director and group data and artificial intelligence officer · Mar 29, 2024
“It is thus essential to adopt an approach in which humans are at the centre of projects and processes.”[15]
Marco Ditta · Head of Data & Artificial Intelligence · Sep 29, 2023
“artificial intelligence, at the end of the day, is software.”[36]
Valerio Cencig · Lead for compliance and digital transformation · Oct 3, 2024
“Banking supervision is an area where it is fair to speak of information overload, with truly massive and exponential content production,”[29]
Walter Chiaradonna · Head of the supervisory strategic steering department · May 16, 2023
“This is a further step towards making our Group an example of best practice in the technological field as well,”[21]
Carlo Messina · Managing Director and Chief Executive Officer, at the Isybank launch · Jun 16, 2023
Who runs AI at Intesa Sanpaolo?
Massimo Proverbio
Chief Data, AI and Technology Officer (heads the Chief Data, A.I. and Technology Officer Governance Area in the July 2026 organisation chart; earlier titled Chief Data, AI, Innovation and Technology Officer)[7][18][6][16]
Gian Maria Gros-Pietro
Chair of the Board of Directors (signed the 2025 annual report letter on technology and AI)[1][8]
Marco Ditta
Executive director and group data and artificial intelligence officer (as named on Intesa Sanpaolo's March 2024 page; Head of Data & Artificial Intelligence in September 2023)[14][15]
Domenico Fileppo
Executive Director, Operational, IT and Cyber Risk Management (Democratic Data Lab on Google Cloud, as named by Google Cloud)[34]
Matteo Feraboli
Group Senior Director, Cyber Security and Anti-Fraud (Global Cyber Defense Center, as named by Microsoft)[35]
Valerio Cencig
Lead for compliance and digital transformation (as introduced in Accenture's October 2024 responsible-AI interview)[36]
Stefano Lucchini
Chief Institutional Affairs and External Communication Officer (his area launched the Lisa and Aptus.AI regulatory tools in 2023)[16][18]
Walter Chiaradonna
Head of the supervisory strategic steering department (led the Lisa machine-learning tool, May 2023)[29]
Marcello Mentini
Executive Head of Group Regulatory Evolution Agenda (Aptus.AI project, October 2023)[16]
Piero Boccassino
Chair of the Anti-Financial Crime Digital Hub, the consortium the bank formed with the Innovation Center and Turin's two universities[8]
Mario Costantini
Chief Executive Officer and General Manager, Neva SGR (the group's venture capital company)[9]
Antonio Valitutti
Chief Executive Officer of Isybank (as named at the June 2023 launch)[21]
Which regulators govern Intesa Sanpaolo's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| ECB | Intesa Sanpaolo is a significant institution under direct ECB banking supervision; the ECB approved its acquisition of control of Monte dei Paschi di Siena in 2026, its 2026-28 priorities cover digitalisation and AI-related risk, its letter on AI-enabled cyber threats bears on the bank's action plan against frontier AI models, and its cloud outsourcing guide applies to the core migration to Google Cloud. The annual report also describes remediation under the ECB guide on risk data aggregation. | SSM supervisory priorities 2026–28 ECB 'Dear CEO' letter on AI-enabled cybersecurity threats (SSM-2026-0301) ECB Guide on outsourcing cloud services |
| EU AI Act | The Group Guidelines for the Use of AI (effective April 2025) are built on the AI Act's risk classes, the annual report describes an AI compliance framework designed for the Act, and the bank responded to consultations on its implementing rules; credit scoring of individuals is high-risk under the Act, and the 2026-2029 plan applies agentic AI to rating assignment and validation. | Regulation (EU) 2024/1689 EBA factsheet: AI Act implications for the EU banking and payments sector Regulation (EU) 2026/1744 (Digital Omnibus on AI) |
| European Union | Italy's data protection authority, which enforces the GDPR, imposed the EUR 17.6 million fine over profiling customers for the Isybank migration, and the plan cites DORA requirements in its supervision of suppliers, which reaches the hyperscaler agreements for cloud and AI services. | GDPR Article 22 DORA (Regulation (EU) 2022/2554) |
| EBA | The EBA's AI adoption work and third-party risk guidelines frame the group's growing dependence on cloud and AI providers, which the annual report lists as a risk. | EBA report: Rising application of AI in EU banking and payments (Sep 2025) EBA/GL/2026/09 (third-party risk) |
| Basel Committee | The annual report describes work to meet supervisory expectations on risk data aggregation and reporting, the subject of BCBS 239, and the move of group data to Google BigQuery and of core systems to Google Cloud engages the Basel third-party risk principles. | BCBS 239 BCBS Third-Party Risk Principles (Dec 2025) |
| Banca d'Italia (Bank of Italy) | The national competent authority that works with the ECB on supervising the group; its anti-money-laundering supervision unit and the Financial Intelligence Unit took part in the 2026 AI tax-fraud pilot with the bank and the Anti-Financial Crime Digital Hub. | |
| Garante per la protezione dei dati personali (Italian Data Protection Authority) | Fined the bank EUR 17.6 million on 12 March 2026 for unlawful processing of the data of about 2.4 million customers profiled and moved to Isybank. |
Side by side: EU AI Act vs Colorado AI Act · all comparisons.
What does the public record suggest about Intesa Sanpaolo's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
Use cases are counted; value is not
Intesa publishes counts (190 developed in March 2025, 153 industrialised by end-2025, 33 launched in 2026) and targets (more than 20% productivity, more than 80% automated resolutions), but no AI value figure. The EUR 570 million of 2030 savings is attributed to generational change made possible by technology and AI together, so the AI share cannot be separated from the public record.
02
Cloud came first, and AI rides on it
The bank spent 2022-2025 rebuilding its core on isytech and Google Cloud and describes the result as AI-ready; agentic AI became a named plan lever only with the 2026-2029 plan. The hyperscaler agreements for 2026-2032 are not named, which leaves the main model and cloud suppliers for the agentic phase undisclosed.
03
The workforce plan is the main AI lever
About 12,400 exits and 6,300 young hires by 2029, after 9,000 exits agreed in 2024 'in the digitalisation and artificial intelligence scenario', make headcount the clearest channel through which AI shows up in the accounts. A bilateral Digital Transformation, AI and Privacy Committee with the unions monitors the effects.
04
Two 2026 incidents show the risk side
The data protection authority's fine concerned profiling of customers for the Isybank migration, and the Fideuram transfers were triggered by an AI-cloned voice. The plan's emphasis on fraud and scam controls, the NEMO personal-data programme and the action plan against frontier AI models are the responses on record; outcomes have not yet been reported.
Which leader rankings does Intesa Sanpaolo appear in?
14 of the computed leaders rankings include Intesa Sanpaolo, each ranked from the same sourced records (bank pages, filings, hiring snapshots, the agent tracker, the incidents register).
How does Intesa Sanpaolo compare with its peers?
- How does Intesa Sanpaolo compare with its peers on AI and technology?
- Intesa Sanpaolo competitors and alternatives, by business line
Its peer group on this site, by business model and size:
How is Intesa Sanpaolo using AI?
Mainly for internal productivity and control: 153 use cases industrialised under its AIxeleration programme by end-2025, a Digital Branch where a virtual agent handles 4 to 5 million contacts a year and a GenAI copilot supports staff, AI in fraud, anti-money-laundering and risk models, and machine-learning tools for regulatory reading. Its 2026-2029 plan adds agentic AI for relationship managers, credit, KYC and AML, and software development.
Who leads AI at Intesa Sanpaolo?
Massimo Proverbio, Chief Data, AI and Technology Officer, runs data, AI and technology and was the bank's spokesman on the move of core IT to Google Cloud; CEO Carlo Messina owns the 2026-2029 plan that makes AI a cost lever. Marco Ditta was named in 2024 as the group's data and artificial intelligence officer, and the Chair, Gian Maria Gros-Pietro, signed the 2025 annual report letter on technology and AI.
Is Intesa Sanpaolo cutting jobs because of AI?
The bank plans about 12,400 exits and 6,300 young hires by 2029, a net cut of about 6,100, and says the related savings of about EUR 570 million come with no revenue impact thanks to technology and AI-enabled streamlining. Exits are voluntary or natural turnover; the 2024 union agreement on 9,000 exits was described as a response to a digitalisation and AI scenario.
What is AIxeleration?
Intesa Sanpaolo's programme, launched in 2023, to scale AI across the group in a standardised and responsible way. By the end of 2025 it had industrialised 153 use cases, 49 of them in 2025, with about 300 dedicated AI specialists, a first MLOps platform, an internal AI Academy and a Data and AI Community of more than 9,000 colleagues.
Does Intesa Sanpaolo use Google Cloud?
Yes. Isybank was built in 2023 on Thought Machine's core in Google Cloud's Italian regions, the risk team runs a Democratic Data Lab on Vertex AI, BigQuery and Gemini, and in July 2026 the bank completed moving its core IT systems, more than 800 applications, to Google Cloud regions in TIM's data centres in Turin and Milan.
What happened in the Fideuram AI scam?
Private Banker International, citing Reuters sources, reported in September 2026 that fraudsters impersonated Intesa Sanpaolo's CEO on WhatsApp and used AI to reproduce a lawyer's voice, leading Fideuram's then chairman to order payments abroad. About EUR 53 million of EUR 95 million was recovered; the bank and Fideuram declined to comment.
Why was Intesa Sanpaolo fined over Isybank?
On 12 March 2026 Italy's data protection authority fined the bank EUR 17.6 million for unlawful processing of the data of about 2.4 million customers it moved to Isybank. The customers were profiled by factors such as being under 65, how often they used digital channels and their investment holdings, and the authority found communication about the migration inadequate.
How does Intesa Sanpaolo govern AI under the EU AI Act?
Its Group Guidelines for the Use of AI, effective April 2025 and approved by the board, follow the AI Act's risk classes: no prohibited practices, the strictest requirements for high-risk systems, human-in-the-loop or human-on-the-loop oversight with the business owner accountable, and guardrails for generative AI. The annual report adds an AI compliance framework and AI metrics in the group's risk appetite.
- Annual Report 2025
- Intesa Sanpaolo: consolidated results as at 30 September 2026 (press release)
- 9M26 Results: The best Q3 and 9M ever
- Intesa Sanpaolo: 2026-2029 Business Plan (press release)
- 2026-2029 Business Plan (presentation)
- Accelerazione su cloud e AI: oltre €10 miliardi in tecnologia entro il 2029
- Intesa Sanpaolo brings its digital infrastructure to Google Cloud regions in Italy hosted in TIM's data centers
- Intesa Sanpaolo and the Anti-Financial Crime Digital Hub: public-private cooperation and artificial intelligence as key tools in the fight against financial crime
- Neva SGR invests in Straiker, a leader in agentic AI security
- Group Guidelines for the Use of Artificial Intelligence: Abstract (effective April 2025; undated page, retrieved 2026-11-19)
- ABI Innovation Awards: Intesa Sanpaolo wins with SofIA and ESG Cash Pooling
- Intesa Sanpaolo joins Frontria consortium to tackle AI-generated disinformation
- IA: una nuova frontiera per la competitività di Intesa Sanpaolo
- How Artificial Intelligence will change banking
- Intesa Sanpaolo chooses iGenius for generative business intelligence and analytics
- Intesa Sanpaolo first European bank to use artificial intelligence for regulatory analysis
- Intesa Sanpaolo accelerates generational change through 9,000 exits at no social cost, with a resilient business model in the digitalisation and artificial intelligence scenario
- Organisation chart of the Central Structures (1 July 2026)
- Statement by Carlo Messina, CEO of Intesa Sanpaolo: results as at 30 September 2026
- Artificial Intelligence Lab (undated page, retrieved 2026-11-19)
- Intesa Sanpaolo launches new digital bank Isybank
- Italy's Intesa to grow profit with cost cuts, return 50 billion euros to investors
- Intesa Plans to Cut 6,100 Jobs by 2029 in Savings Push
- Italy fines Intesa Sanpaolo 18 mln euros for illicit processing of customer data
- Intesa’s private banking arm hit by €95m voice AI scam – report
- Italian bank Intesa strikes deal with unions to cut staff costs
- AI Push Sees Intesa Axing 9,000 Jobs in ‘Generational’ Shift
- Intesa launches cloud-based Isybank in digital push
- Italy's Intesa embraces AI to tame bank supervision "information overload"
- RPT-BRIEF-Intesa Sanpaolo Signs MoU With Tim And Google Cloud For Bank's Digitalization
- Intesa Sanpaolo migrates core IT systems to Google Cloud
- Swift and top banks test use of AI and data sharing to fight financial fraud
- Earnings call transcript: Intesa Sanpaolo Q3 2026 profit tops 8.2 billion euros
- Intesa Sanpaolo: Managing evolving financial risk at speed with data analytics and AI (undated page, retrieved 2026-11-19)
- Intesa Sanpaolo accrues big cybersecurity dividends with Microsoft Sentinel, Copilot for Security (undated page, retrieved 2026-11-19)
- Rethinking responsible AI: a conversation with Valerio Cencig, Intesa Sanpaolo (video transcript)
- Intesa Sanpaolo’s Isybank goes live on Thought Machine's Vault Core
- Evident AI Index | Banks (undated page, retrieved 2026-11-19)
When Intesa Sanpaolo 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 PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning