'AI for everyone, everywhere, everything': one platform, half the staff building agents, 'digital employees' on the org chart.
How is BNY using AI?
BNY, America's oldest bank and the world's largest custodian, spends a larger share of revenue on technology than any big-bank peer — about $3.8 billion in 2025, roughly 19% of revenue — and CEO Robin Vince's mantra is that 'AI is for everyone, everywhere and for everything'. The vehicle is Eliza, an enterprise platform that nearly every employee is trained on and that more than half of them now use to build agents; the bank counted 117 AI solutions in production at September 30, 2025, about 220 (alongside 140 'digital employees') in April 2026 and '400-plus' by Vince's count on the October 15, 2026 earnings call. BNY has tripled its intern and analyst classes since 2022 to hire for AI proficiency, saw headcount fall 7% year on year by mid-2026, and in July 2026 Vince joined OpenAI's board — the clearest public tie between a systemically important bank and a frontier-model developer.
Lead bank
The Bank of New York Mellon
Assets (lead bank, Jun 30, 2026)
$425 billion
Charter
State member bank (Fed)
Headquarters
New York, NY
Flagship platform
Eliza
AI leadership
Robin Vince · Leigh-Ann Russell · Sarthak Pattanaik · Michael Demissie · Carolyn Weinberg · Jennifer Barker · Carl Slabicki · Dermot McDonogh · Nastassja Hagan · Watt Wanapha · Michelle O'Reilly · Rachel Lewis · Shannon Hobbs · Ainslie Simmonds · Ben Harrison · Victor O'Laughlen · Bridget Engle
- Technology spend of about $3.8 billion in 2025, roughly 19% of revenue — the highest proportion among large US banks, per CNBC's comparison.
- Eliza platform: nearly all employees trained; ~50% daily AI users and >50% building agents; '400-plus' AI solutions in production (CEO Robin Vince, Q3 2026 call, October 15, 2026), up from ~220 solutions and ~140 digital employees in the Q1 2026 deck and 117 solutions at September 30, 2025.
- Model partners on the platform: a multi-year OpenAI deal (February 2025, Deep Research and reasoning models) and Google Cloud's Gemini Enterprise (December 2025) for employee-built data-analysis agents.
- Workforce: intern and analyst classes tripled since 2022 to tap AI proficiency; headcount down 7% year on year in Q2 2026 while revenue hit a record $5.7 billion; Vince calls AI 'a jobs creator, not a destroyer'.
- Research and outreach: a five-year, $10 million AI R&D agreement with Carnegie Mellon (September 2025); a Future of Work Alliance with the University of Manchester (May 2026); free AI and cyber training for 1,000 community-bank executives.
- Depth of use (CFO, July 2026): more than 40% of code AI-authored in the first quarter of 2026 and about 50% since; about half of annual account plans drafted with AI, 25% of client onboarding AI-supported and roughly 70% of restricted-party payment screening reviewed by AI.
- Governance on the record: the 2025 10-K says model risk management covers AI use cases including large language models inside a cross-functional AI risk framework; digital employees are 'onboarded, governed, and continuously monitored', says CIO Leigh-Ann Russell.
- July 2026: CEO Robin Vince appointed to OpenAI's for-profit and nonprofit boards ahead of its IPO.
- Infrastructure: in March 2024 BNY became the first major bank to deploy an NVIDIA DGX SuperPOD (DGX H100), then supporting deposit forecasting, payment automation and predictive trade analytics; the bank says the on-premises GPUs supplement its cloud capacity, and its chief data and AI officer said in 2025 that Anthropic, Google and OpenAI models run alongside Llama and Mistral models that BNY hosts on premises.
- Adoption curve: about 14,000 staff (a quarter of the workforce) used Eliza in August 2024; American Banker reports about 20% usage at the start of 2025, 99% of employees trained by mid-2025 and more than half now building agents; OpenAI's case study counts 20,000 employees building agents and more than 125 live use cases in December 2025.
- Governance path: a 16-step review route from idea to working agent (Microsoft, May 2026), with a data use review board, an AI release board and an Enterprise AI Council (OpenAI case study); Eliza applies permissions, tagging and telemetry to every model and agent.
- Measured outputs in the Q1 2026 deck: more than 80% faster investigation of custody settlement inquiries handled by AI, more than 30% faster restricted-party screening, more than 10% more software releases and 171,000 AI learning hours delivered in 2025.
- Cost framing: on the Q2 2026 call the CFO called BNY's AI spend 'de minimis' within an engineering budget of about $4 billion; the Q2 deck says the platform operating model finished its activation phase and the bank is tracking toward about $500 million of incremental investment and about $450 million of efficiency savings in 2026.
- Cyber and frontier access: in April 2026 BNY was among the firms given early access to OpenAI's GPT-5.4-Cyber and Anthropic's Mythos models, and its Q2 2026 10-Q says recent advances in AI have enhanced threat actors' ability to find and exploit previously unidentified vulnerabilities.
What is Eliza?
BNY's enterprise AI platform, on which nearly all employees are trained; Eliza 2.0 launched in September 2025, OpenAI's models and Deep Research were added under a multi-year deal in February 2025 and Google Cloud's Gemini Enterprise in December 2025. By the first quarter of 2026 about 50% of employees were daily AI users, more than 50% were building agents, and the bank counted roughly 220 enterprise AI solutions and 140 'digital employees' in production; on the Q3 2026 earnings call (October 15, 2026) CEO Robin Vince put the number of AI solutions in production at '400-plus'.[14][1][3][8][21][17]
What AI partnerships and patents does BNY have?
BNY runs a model-agnostic platform with OpenAI, Google and Anthropic models, hosts open-weight models on premises, and was the first major bank to deploy an NVIDIA DGX SuperPOD (March 2024). Microsoft supplies Azure and Copilot and BNY is an inaugural member of its Frontier Firm AI Initiative; it also joined a federated-learning fraud proof of concept with Kinexys, RBC, DeepTempo and NVIDIA. Its Q4 2025 deck counted more than 60 AI-related patent applications, and public records show a granted time-series patent plus published applications on LLM classification, agent orchestration, handwriting extraction and code translation.[25][42][29][49][39][56][54][57]
| Item | What it is | Status | Date |
|---|---|---|---|
| OpenAI | Multi-year collaboration on Deep Research and reasoning models for Eliza; CEO Robin Vince joined OpenAI's boards in July 2026 | In production | 2025-02-26[8][38][11] |
| Google Cloud | Gemini Enterprise integrated into Eliza for agentic deep research; Gemini 3 and Veo 3 also in use | In production | 2025-12-08[46] |
| Anthropic | Claude models on Eliza (Q1 2026 deck); BNY named in Anthropic's finance-agent launch; early access to Mythos cyber model | In production | 2026-05-05[39][50][48] |
| Microsoft | Azure cloud and Copilot for staff; inaugural member of the Frontier Firm AI Initiative with the Harvard Business School AI Institute | In production | 2026-05-12[29] |
| NVIDIA | DGX SuperPOD (DGX H100) deployed March 2024; NVIDIA FLARE and DGX Cloud used in the 2026 federated-learning fraud proof of concept | In production | 2024-03-18[42][49] |
| Kinexys by J.P. Morgan, RBC, DeepTempo | Federated-learning fraud detection proof of concept (Project Aikya) | Pilot | 2026-04-27[49] |
| Behavox | Quantum AI for communications surveillance across 16-plus languages | In production | 2025-04-17[24] |
| Quantexa | Strategic investment in the entity-resolution and graph analytics vendor | Invested | 2021-09-30[35] |
| Carnegie Mellon University | Five-year, $10 million AI research and development agreement | Active | 2025-09-11[6] |
| University of Manchester | Future of Work Alliance on responsible, human-led AI (five years, GBP 5 million) | Active | 2026-05-21[7] |
| US 2025/0217603 A1 | Large language model and neural networks for classifying concepts in natural language text | Published application | 2025-07-03[54] |
| US 2025/0182530 A1 | Template-independent handwriting extraction from images with machine-learning models | Published application | 2025-06-05[55] |
| US 12,417,410 B2 | Real-time optimal data predictions for time series signals (model selection, noise filtering, window size) | Granted | 2025-09-16[56] |
| US 2026/0057262 A1 | Orchestrating interaction with an AI application through an AI agent and an API repository | Published application | 2026-02-26[57] |
| EP 4758506 A1 | Machine-learning translation of one coding language into another | Published application | 2026-06-17[58] |
What has BNY done on AI, and when?
- Oct 15, 2026
Q3 call: '400-plus' AI solutions in production; AI spend not broken out[14]
Vince says the count of AI solutions in production, 117 a year earlier, is now '400-plus'; he declines to give an AI dollar figure ('we're not breaking it out'), says about $4 billion of the $13 billion annual expense base goes on technology and that 'yes, we're generating returns, and yes, we want to lean in more'.
- Oct 1, 2026
Russell profile: usage up from about 20% to 99% trained in a year[53]
American Banker says about 20% of staff used Eliza at the start of 2025 against a 65% year-end target; 99% were trained within six months, more than 50% build agents and over 60% of code is built with AI help. It also credits Russell with digital employees that detect and fix code vulnerabilities, now more than 150.
- Sep 19, 2026
CFO: ~$4 billion annual engineering spend funds Eliza; expense outlook raised partly on AI[37]
At an investor conference McDonogh frames BNY's next growth phase around cross-selling and AI investment, says AI is meant to create capacity to reinvest in client delivery and product development rather than just cut headcount, flags AI's potential to modernise spreadsheet-heavy loan-market processes, and raises the 2026 expense-growth outlook to 6%-7%, partly on added data and AI investment.
- Sep 17, 2026
AI named as a factor amplifying payments fraud[36]
Carl Slabicki, Head of Commercial, Global Payments & Trade, tells PYMNTS that AI is changing the fraud landscape by amplifying social engineering and making deepfakes and synthetic-identity fraud more convincing.
- Aug 20, 2026
AI behind payments facilitation[33]
Payments head Jennifer Barker, who has led several AI initiatives, describes intelligent routing that picks the payment rail from client preferences.
- Aug 12, 2026
Vince rates his own AI use nine out of 10[52]
In an Axios interview two weeks after joining OpenAI's board, Vince describes an Eliza daily planner that reads his calendar, email and documents, a link between Eliza and the 360 feedback process, and model routing chosen by the app; he cites 46,000 employees.
- Jul 31, 2026
Q2 10-Q: AI enhances cyber threats[41]
The quarterly report says recent advancements in artificial intelligence have enhanced threat actors' capability to identify and potentially exploit previously unidentified cybersecurity vulnerabilities, and lists AI among the forward-looking topics alongside the transition to a platforms operating model.
- Jul 28, 2026
- Jul 21, 2026
- Jul 15, 2026
Q2 call and deck: AI value in three dimensions; operating model activated[30][40]
Vince frames AI as running the company better, building better products and expanding BNY's perimeter, and says it is 'a mistake' to see it only as efficiency. The deck says the platform operating model's activation phase finished in Q2 and BNY is tracking toward about $500 million of incremental investment and about $450 million of efficiency savings in 2026; the CFO puts AI spend as 'de minimis' within a roughly $4 billion engineering budget.
- Jul 15, 2026
- Jun 4, 2026
INSITE 2026: AI in wealth workflows[51]
Vince says 'What AI is going to do is give us all superpowers' and that BNY spends about $4 billion a year on technology; Ainslie Simmonds says BNY has automated part of account transfers and points to account opening, transfers and KYC as workflows where AI helps; BNY Pershing oversees about $3.3 trillion of client assets.
- May 28, 2026
Intern and analyst classes tripled since 2022[13]
Vince, 'an AI optimist', hires young staff for their AI proficiency.
- May 21, 2026
Future of Work Alliance with the University of Manchester[7]
Responsible, human-led AI research and testing.
- May 12, 2026
Inside the digital-employee programme[29]
Microsoft profile: about 2,300 staff through a 40-hour AI bootcamp; a payments digital employee cut open investigations by nearly 80%; a 'Digital Employee Agency' decides when to deploy agents rather than hire.
- May 5, 2026
Named among Anthropic's finance-agent customers[50]
Anthropic's launch of ten finance agent templates carries a BNY statement: 'With Eliza and Claude, we're giving processes new digital employees who work the case end to end.'
- May 4, 2026
Vince: AI is 'a jobs creator, not a destroyer'[19]
At Milken: savings in one place create capacity to do more elsewhere.
- Apr 27, 2026
Federated-learning fraud detection proof of concept[49]
Kinexys by J.P. Morgan, BNY, RBC, DeepTempo and NVIDIA trained local fraud models without pooling transaction data, using NVIDIA FLARE on DGX Cloud; the result outperformed single-institution models and nearly matched a centralised one.
- Apr 16, 2026
Early access to OpenAI and Anthropic cyber models[48]
Axios reports BNY is among the firms testing GPT-5.4-Cyber and Anthropic's Mythos after Treasury Secretary Bessent and Fed Chair Powell convened bank CEOs on AI-enabled attack risk; Vince says the labs want real-world feedback.
- Apr 16, 2026
Q1 call: 218 AI solutions in production, up fourfold[28]
Vince recounts the 2023 AI Hub and BNY's DGX SuperPOD with NVIDIA, the first at a global bank, and says 2026 moves from point solutions to end-to-end processes.
- Apr 16, 2026
- Mar 11, 2026
- Feb 26, 2026
Patent application: orchestrating interaction with an AI application[57]
US 2026/0057262 A1 describes an AI agent that turns a user message into an inference process, checks an API repository for missing inputs and then executes it.
- Feb 25, 2026
10-K: AI inside model risk management[27]
Model risk management covers AI use cases including large language models and sits within a cross-functional AI risk framework; a dedicated risk factor covers the development and use of AI.
- Feb 9, 2026
134 'digital employees'; $3.8 billion tech spend[10]
CNBC: about 19% of revenue, the highest proportion among peers.
- Jan 13, 2026
- Dec 12, 2025
OpenAI case study: 20,000 agent builders, 125+ live use cases[38]
Pattanaik says 'We're not building side projects. We're changing how the bank works.' A Contract Review Assistant cut legal review from four hours to one across 3,000+ annual vendor agreements, and 99% of staff were trained on generative AI.
- Dec 8, 2025
- Oct 28, 2025
Vince at Money20/20: adoption, not capability, is the constraint[20]
'Human adoption and culture and potentially organizational inertia' will get in the way.
- Oct 20, 2025
- Oct 16, 2025
117 AI solutions in production; Eliza 2.0[17]
Up 75% from the second quarter; Vince: 'AI is for everyone, everywhere and for everything'.
- Sep 29, 2025
Eliza 2.0 shown to staff: 98% trained, majority daily users[45]
Vince unveiled the next Eliza version at an internal town hall; an email from Chief People Officer Shannon Hobbs and CIO Leigh-Ann Russell said it would cut client briefings from hours to minutes and that 'Fluency in AI is no longer optional, it's critical'.
- Sep 11, 2025
$10 million, five-year AI R&D agreement with Carnegie Mellon[6]
Students, faculty and BNY experts on AI applications and systems.
- Sep 10, 2025
CIO: among the 5% seeing meaningful AI ROI[9]
Russell on Eliza and training employees toward 100% AI adoption.
- Feb 26, 2025
Multi-year OpenAI deal[8]
Access to Deep Research and OpenAI's most advanced reasoning models to 'supercharge' Eliza.
- Jan 22, 2025
Enterprise gen-AI assistant and anomaly detection described[5]
An AI solution for anomaly detection in daily calculations cuts hours to minutes; a firm-wide generative assistant helps staff find procedures.
- Aug 15, 2024
Eliza described: about 14,000 users, built on GPT-4, Gemini and Llama[43]
Fortune reports Eliza, then used by a quarter of the workforce, runs on Nvidia chips with Azure and Google Cloud and draws on OpenAI's GPT-4, Google's Gemini and Meta's Llama. It traces the platform to summer 2023, when Robin Vince asked Sarthak Pattanaik to fold scattered AI projects, some taking 18 months to build, into one central platform.
- Mar 18, 2024
First major bank to deploy an NVIDIA DGX SuperPOD[42]
BNY revealed a SuperPOD built on DGX H100 systems with InfiniBand networking, installed ahead of typical timelines; then-CIO Bridget Engle said it would accelerate processing capacity. The AI Hub had more than 20 solutions in production, a firmwide exercise had found more than 600 AI opportunities, and early use cases were deposit forecasting, payment automation, predictive trade analytics and end-of-day cash balances.
- Sep 30, 2021
Strategic investment in Quantexa[35]
BNY Mellon joined Quantexa's Series D after a 12-month engagement using its platform, with an expanded relationship focused on 'data fabric' innovation (master data management, entity resolution, graph network intelligence and AI); its CIO of engineering and architecture joined Quantexa's board and its chief data officer the advisory board.
Where does BNY use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| Eliza and employee-built agents[1][10] | Enterprise platform; more than half of employees building agents; ~140 digital employees on repetitive tasks. | In production | Generative & agentic AI |
| Anomaly detection in daily calculations[5] | Flags areas for review in minutes instead of hours. | In production | AI governance (general) |
| AI-generated code[18] | The CFO says engineers write a growing share of code with AI; the bank routes tasks to the most appropriate models. | In production | Generative & agentic AI |
| OpenAI and Google Cloud models on Eliza[8][21] | Multi-model platform: OpenAI (Feb 2025) and Gemini Enterprise (Dec 2025). | In production | Third-party & vendor AI |
| Community-bank AI training[23] | Free AI and cyber training for 1,000 community-bank executives, led by senior BNY leaders. | In production | AI governance (general) |
| Payments digital employees[29] | Agents that repair cross-border payment data and answer client payment inquiries; validations down from five or six minutes to under 30 seconds. | In production | Generative & agentic AI |
| AI review of restricted-party screening and communications surveillance[31][24] | Roughly 70% of restricted-party payment screening reviewed by AI; Behavox Quantum AI covers BNY's communication channels across 16-plus languages for compliance and conduct. | In production | AML / KYC |
| Azure document intelligence in lockbox[29] | More than 500,000 check images a day read with Azure document intelligence, alongside Copilot tools for employees. | In production | Third-party & vendor AI |
| Model risk management for AI and LLMs[27] | An active model inventory and independent validation extended to AI use cases, including large language models, within a cross-functional AI risk framework. | In production | Model risk management |
| Contract Review Assistant[38] | Agent built on Eliza that cut legal review of vendor agreements from four hours to one (75%) across more than 3,000 agreements a year. | In production | Generative & agentic AI |
| Settlement-inquiry digital employee[39] | More than 10% of custody settlement status inquiries are addressed by AI as of Q1 2026, with investigations more than 80% faster than a year earlier. | In production | Generative & agentic AI |
| Frontier cyber-model testing[48][53] | Early access to OpenAI GPT-5.4-Cyber and Anthropic Mythos for defensive testing; digital employees also find and fix code vulnerabilities. | Pilot | Cybersecurity |
| Federated-learning fraud detection[49] | Proof of concept with Kinexys by J.P. Morgan, RBC, DeepTempo and NVIDIA: banks train local models and share encrypted updates instead of transaction data. | Pilot | Fraud detection |
| Walled-garden training for sensitive data[44] | BNY avoids training models on personal information; where it does, it uses an encrypted 'walled garden' with red-team testing. | In production | Data & privacy |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
Which AI vendors and fintechs does BNY work with?
| Company | What BNY uses it for | Status | Category |
|---|---|---|---|
| Quantexa | Data unification, entity resolution, master data management and graph analytics ('data fabric') across the bank since a 12-month engagement before its 2021 investment; an early adopter of the Q Assist generative AI copilot in 2024. Not described as an AML deployment. | In production | AML compliance |
| Snapdocs | Joint initiative, announced May 2026, to deliver automated collateral delivery and eCustody that combines BNY's document custody with Snapdocs' eVault, document classification and data extraction; expanded October 16, 2026 into VeriCORE, a collateral ownership and control registry for consumer lending and ABS, founded jointly with the Structured Finance Association and targeting an initial 2027 launch. | Announced | Document AI |
| Microsoft 365 Copilot | Copilot made broadly available to employees in November 2023; roughly half the workforce is licensed for Microsoft 365 Copilot, per Microsoft WorkLab. | In production | Enterprise AI assistants |
| ChatGPT Enterprise (OpenAI) | ChatGPT and the API inside BNY's Eliza platform, with OpenAI frontier models; OpenAI reports 20,000 employees building agents. | In production | Enterprise AI assistants |
| Claude (Anthropic) | Claude models run on BNY's Eliza platform; Anthropic's launch of ten finance-agent templates carries a BNY statement ('With Eliza and Claude, we're giving processes new digital employees who work the case end to end'); also among firms given early access to the restricted Claude Mythos cyber model (April 2026). | In production | Enterprise AI assistants |
| Gemini Enterprise (Google Cloud) | Gemini Enterprise integrated into BNY's Eliza platform so employees can build data-analysis agents; Google lists BNY among financial institutions using it. | In production | Enterprise AI assistants |
Each vendor and fintech page cites the source for the relationship, and vendor pages list products, pricing and other bank customers. Compare AI tools for banks by category, or read the fintech and neobank AI dossiers.
What has BNY disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Technology spend | ≈ $3.8 billion (2025), ≈ 19% of revenue[10] | Feb 9, 2026 |
| Daily AI users / agent builders | ≈ 50% / > 50% of employees[1] | Apr 16, 2026 |
| AI solutions in production | 400-plus (CEO, Q3 2026 call)[14] | Oct 15, 2026 |
| AI solutions / digital employees in production | ≈ 220 / ≈ 140[1] | Apr 16, 2026 |
| Headcount change | −7% year on year[18] | Jul 15, 2026 |
| Assets under custody / administration | $63.4 trillion (preliminary)[15] | Sep 30, 2026 |
| Carnegie Mellon AI agreement | $10 million over five years[6] | Sep 11, 2025 |
| Employees on Eliza / AI solutions / digital employees / AI patent applications | 99% / 160 / 134 / > 60[25] | Dec 31, 2025 |
| Code authored or written with AI | > 40% in Q1 2026 (CFO); ≈ 50% by July 2026 (Fortune); > 60% built with AI help by October 2026 (American Banker)[30][31][53] | Oct 1, 2026 |
| Account plans drafted with AI / onboarding AI-supported / restricted-party screening reviewed by AI | ≈ 50% / 25% / ≈ 70%[31] | Jul 28, 2026 |
| Revenue per employee | $338,000 (2022) → $401,000 (2025)[31] | Dec 31, 2025 |
| AI bootcamp graduates (40-hour programme) | ≈ 2,300 since April 2025[29] | May 12, 2026 |
| Annual engineering spend (funds the Eliza platform) | ≈ $4 billion[37] | Sep 19, 2026 |
| 2026 expense-growth outlook | 6%-7% (raised, partly on data/AI investment)[37] | Sep 19, 2026 |
| Employees building agents / live use cases (OpenAI case study) | ≈ 20,000 / > 125[38] | Dec 12, 2025 |
| AI learning hours delivered in 2025 | 171,000[39] | Apr 16, 2026 |
| Settlement inquiries addressed by AI / investigation speed-up | > 10% / > 80% faster (1Q26 vs 1Q25)[39] | Apr 16, 2026 |
| Contract review time, vendor agreements | 4 hours to 1 hour (-75%), > 3,000 agreements a year[38] | Dec 12, 2025 |
| 2026 incremental investment / efficiency savings (tracking) | ≈ $500 million / ≈ $450 million[40] | Jul 15, 2026 |
| Eliza reach: Aug 2024 users vs mid-2025 trained | ≈ 14,000 (25% of staff) vs 99% trained[43][53] | Oct 1, 2026 |
| Headcount | 48,100 (53,400 in 2023)[10] | Feb 9, 2026 |
“At BNY, AI is for everyone, everywhere and for everything.”[17]
Robin Vince · Chairman and CEO · Oct 16, 2025
“I am an AI optimist. I think this is a revolutionary technology that for sure has to be well-managed.”[13]
Robin Vince · Chairman and CEO · May 28, 2026
“We've been quite constrained in our spending, and we've been very disciplined, we continue to spend more on cyber resiliency than we do on AI.”[26]
Dermot McDonogh · Chief Financial Officer · Jan 13, 2026
“At BNY, we continue to view AI as 1 of the most important long term opportunities for our company.”[30]
Robin Vince · Chairman and CEO · Jul 15, 2026
“We launched our first digital employee in 2025 with the same approach as our enterprise-wide systems—governance was built in from the start.”[32]
Leigh-Ann Russell · CIO and Global Head of Engineering · Aug 4, 2026
“AI is changing the fraud landscape, amplifying social engineering, making deepfakes convincing, and scaling synthetic identity and account takeover.”[36]
Carl Slabicki · Head of Commercial, Global Payments & Trade · Sep 17, 2026
“In 2026, we are doubling down on depth, moving from AI point solutions to using AI to enhance end-to-end processes, reducing manual touch points, improving cycle times, strengthening control outcomes”[28]
Robin Vince · Chairman and CEO · Apr 16, 2026
“It would be a mistake to think about this as just an efficiency creator.”[30]
Robin Vince · Chairman and CEO · Jul 15, 2026
“within the context of $4 billion our AI spend quite de minimis and quite modest and appropriate for the strategy that we have in place.”[30]
Dermot McDonogh · Chief Financial Officer · Jul 15, 2026
“We're not building side projects. We're changing how the bank works.”[38]
Sarthak Pattanaik · Chief Data and AI Officer · Dec 12, 2025
“Some might see AI governance as a barrier, but in our experience, it's been an enabler.”[38]
Watt Wanapha · Deputy General Counsel and Chief Technology Counsel · Dec 12, 2025
Who runs AI at BNY?
Robin VinceLinkedIn ↗
Chairman and CEO; OpenAI board member since July 2026[12][20]
Sarthak Pattanaik
Chief Data and AI Officer; led the 2023 consolidation of AI work into the AI Hub and the Eliza platform[21][43][44][46][38]
Michael DemissieLinkedIn ↗
Head of Applied AI[29]
Jennifer BarkerLinkedIn ↗
Global Head of Payments & Trade and Depositary Receipts[33]
Carl Slabicki
Head of Commercial, Global Payments & Trade[36]
Dermot McDonoghLinkedIn ↗
Nastassja Hagan
Head of Applied AI; joined BNY in 2025 from BP, where she worked with CIO Leigh-Ann Russell[53]
Watt Wanapha
Deputy General Counsel and Chief Technology Counsel; sits on AI use-case governance and uses deep research daily[38]
Michelle O'Reilly
Global Head of Talent; leads the AI bootcamps and the planned Digital Employee Agency[38][29][10]
Rachel Lewis
Head of payment operations; leads AI enablement across operations and oversees nine digital employees[10][29]
Shannon Hobbs
Chief People Officer; co-signed the Eliza 2.0 launch email to staff with the CIO[45]
Ainslie Simmonds
Executive platform owner, wealth services platform; applies AI to account opening, transfers and KYC[51]
Ben Harrison
Managing Director, BNY Pershing; reported 160 'digital employees' in March 2026[47]
Victor O'Laughlen
Managing Director, head of the global clearance platform (2024); early Eliza agent builder[43]
Bridget Engle
Chief Information Officer at the time of the March 2024 DGX SuperPOD deployment[42]
What AI roles is BNY hiring for?
Data as of · 12 AI-titled postings · read from BNY's own careers site, re-checked weekly
12 open AI-related postings on BNY's careers site as of Oct 25, 2026. A snapshot of current openings, not a source: postings close. Compare with other banks on which banks are hiring for AI.
Senior Director, AI / Machine Learning Data Engineer
New York, NY, United States · posted
Director, Artificial Intelligence Risk Management
London, United Kingdom · posted
Senior Vice President, AI Governance & Controls (APAC AI Hub)
Singapore, Singapore, Singapore · posted
Director- AI Security Product Manager
Pittsburgh, PA, United States · posted
Director, Applied AI Product Manager
New York, NY, United States · posted
Director, AI / Machine Learning Software Engineer
Jersey City, NJ, United States · posted
Director, AI / Machine Learning Data Engineer
Lake Mary, FL, United States · posted
SVP - Model Risk Management AI, Wealth and Investments
New York, NY, United States · posted
Senior Vice President, Data Science Manager
Boston, MA, United States · posted
Senior Vice President, AI / Machine Learning Software Engineer
Jersey City, NJ, United States · posted
Senior Vice President, AI / Machine Learning Data Engineer
Jersey City, NJ, United States · posted
Pittsburgh, PA, United States · posted
Which regulators govern BNY's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| Federal Reserve | State member bank and G-SIB holding company; SR 26-2 and the third-party guidance govern the OpenAI and Google dependencies on Eliza. | SR 26-2 SR 23-4 |
| NY DFS | New York-chartered: Part 500 certification and the DFS frontier-AI letter apply. | 23 NYCRR Part 500 DFS Frontier AI Models Industry Letter (May 2026) |
| Basel Committee | Custody at $63 trillion makes BCBS 239 data lineage and the third-party principles central. | BCBS 239 BCBS Third-Party Risk Principles (Dec 2025) |
| FSB | G-SIB; the FSB's concentration warning names a few model providers — one of whose boards the CEO now sits on. | FSB AI sound practices consultation (June 2026) FSB AI financial stability report (Nov 2024) |
| SEC | Investment-services and asset-management businesses under SEC exam priorities on AI. | Division of Examinations FY2026 Priorities |
Side by side: SR 11-7 vs SR 26-2 · PRA SS1/23 vs SR 26-2 · MAS AI Risk Management Guidelines vs SR 26-2 · all comparisons.
What does the public record suggest about BNY's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
The most decentralised agent program in banking
More than half of employees building agents is a different governance problem from a central AI team shipping use cases. The 'digital employee' count and the routing of tasks to appropriate models are how the bank makes that legible; supervisors will ask how it is inventoried.
02
A board seat at a model provider is a conflict to be managed, not hidden
Vince on OpenAI's boards while BNY runs OpenAI models on Eliza is disclosed and public. Expect questions on vendor selection and on how the bank's third-party risk program treats the relationship.
03
Headcount is falling while hiring rises — the BNY version of the workforce story
−7% headcount, tripled junior classes and 'jobs creator' rhetoric describe a reshaped pyramid, not a smaller one. It is the model other custody and operations-heavy banks will be measured against.
04
Outcomes over tokens
The CFO's refusal to report prompt counts sets BNY apart from peers reporting prompts per day; it is a deliberate choice that will age well if AI costs keep rising.
Where does the BNY record go deeper?
4 topic pages under this profile, each answering one question from the same sourced record.
Which leader rankings does BNY appear in?
13 of the computed leaders rankings include BNY, each ranked from the same sourced bank pages.
How does BNY compare with its peers?
- How does BNY (Bank of New York Mellon) compare with its peers on AI and technology?
- BNY (Bank of New York Mellon) competitors and alternatives, by business line
Its peer group on this site, by business model and size:
What is Eliza at BNY?
Eliza is BNY's enterprise AI platform, on which nearly all employees are trained. Eliza 2.0 launched in September 2025; it runs OpenAI's models under a multi-year agreement announced in February 2025 and added Google Cloud's Gemini Enterprise in December 2025 so employees can build their own agents. By the first quarter of 2026 about half of employees used AI daily, more than half were building agents, and the bank counted roughly 220 AI solutions and 140 'digital employees' in production. On the Q3 2026 earnings call on October 15, 2026, Vince put the number of AI solutions in production at '400-plus'.
How much does BNY spend on technology?
About $3.8 billion in 2025, roughly 19% of revenue — the highest proportion among large US banks according to CNBC's comparison in February 2026.
Who leads AI at BNY?
CEO Robin Vince sets the 'AI for everyone, everywhere, and everything' agenda and joined OpenAI's boards in July 2026. Sarthak Pattanaik, chief data and AI officer, has run the AI Hub and Eliza platform since 2023. Leigh-Ann Russell, CIO and global head of engineering, drives adoption and digital employees. Michael Demissie and Nastassja Hagan are named as heads of applied AI, Michelle O'Reilly (talent) runs bootcamps and the Digital Employee Agency, Watt Wanapha (technology counsel) sits on governance, and CFO Dermot McDonogh reports outcomes rather than prompt counts.
Which AI models and vendors does BNY use?
Eliza is model agnostic. The Q1 2026 deck lists models from OpenAI, Google and Anthropic, and BNY's chief data and AI officer has said it also hosts Meta's Llama and Mistral models on premises. Partners include OpenAI (multi-year collaboration, February 2025), Google Cloud (Gemini Enterprise, December 2025), Microsoft (Azure, Copilot and the Frontier Firm AI Initiative), NVIDIA (DGX SuperPOD, March 2024) and Behavox for communications surveillance. In April 2026 it also gained early access to OpenAI and Anthropic cyber models.
How does BNY govern its AI agents and 'digital employees'?
BNY routes agents through a 16-step review path of boards, compliance gates and model risk evaluations, according to Microsoft. Cross-functional bodies include a data use review board, an AI release board and an Enterprise AI Council. Each digital employee has login credentials, an employee number and a human supervisor who tracks volume, cycle time and approval outcomes on a dashboard. The 2025 10-K places AI inside model risk management, and the Q2 2026 10-Q flags AI-enhanced cyber threats. Reported gains, such as nearly 80% fewer open payment investigations, are the bank's own figures.
- Quarterly Update Presentation, 1Q 2026
- Quarterly Update Presentation, 2Q 2026
- Unlocking Value with BNY's Enterprise AI Platform
- AI in the Next Phase: Driving Adoption
- Unlocking AI Growth in Financial Services
- BNY and Carnegie Mellon University Join Forces to Advance Research and Development in AI
- BNY and The University of Manchester Launch the Future of Work Alliance
- BNY, America's Oldest Bank, Signs Multiyear Deal With OpenAI
- BNY among 5% seeing meaningful return on investment from AI, says CIO
- Digital employees, AI bootcamps: America's oldest bank is spending billions on tech
- OpenAI appoints two new members to board of directors
- OpenAI Names BNY, Nubank CEOs to Board Ahead of IPO
- BNY Triples Hiring of Interns and Analysts to Tap AI Proficiency
- While embedding AI 'everywhere,' BNY won't disclose costs
- Exhibit 99.1: Earnings Release, Third Quarter 2026 (Form 8-K)
- BNY's profit rises on strong fees, interest income growth
- BNY moves quickly to adopt AI solutions, platforms
- BNY: AI bet is paying off, client impact is 'measurable'
- BNY CEO says AI is a jobs creator, not a destroyer
- BNY's Robin Vince stays big on AI
- BNY looks to Google Cloud for agentic AI push
- AI adoption will trim banking industry costs by up to 20%
- 'These players matter': BNY embraces role on small lenders' AI journey
- BNY Upgrades Compliance Monitoring by Completing Implementation of Behavox Quantum AI
- Exhibit 99.3: Quarterly Update Presentation, 4Q 2025 (Form 8-K)
- BNY Mellon (BK) Q4 2025 Earnings Call Transcript
- Annual Report on Form 10-K for 2025, Exhibit 13.1 (financial section and risk factors)
- BNY Mellon (BK) Q1 2026 Earnings Call Transcript
- The making of a Frontier Firm: How AI is redesigning work at BNY
- BNY (BNY) Q2 2026 Earnings Call Transcript
- BNY skipped the 'tokenmaxxing' craze. Here's what AI metrics it tracks instead
- Your AI agent can be a teammate. But it still needs a boss
- BNY's Barker on what really concerns clients about AI and payments
- BNY's Carolyn Weinberg on trusted infrastructure and where tokenization gets real
- BNY Mellon Invests in Quantexa Technology
- Digital Fraud Forces Issuers to Rethink Trust in the AI Era
- BNY Sees Cross-Selling and AI Investments Fueling Its Next Growth Phase
- BNY builds 'AI for everyone, everywhere' with OpenAI
- Exhibit 99.3: Quarterly Update Presentation, 1Q 2026 (Form 8-K)
- Exhibit 99.3: Quarterly Update Presentation, 2Q 2026 (Form 8-K)
- Quarterly Report on Form 10-Q for the period ended June 30, 2026
- BNY First Global Bank to Deploy AI Supercomputer Powered by NVIDIA DGX SuperPOD With DGX H100
- Exclusive: How BNY's new AI tool Eliza is minting an army of disposable assistants
- Banking on AI: Firms such as BNY balance high risk with the potential for transformative tech
- BNY unveils updated Eliza platform, says 98% of employees are trained on AI
- BNY Collaborates with Google Cloud to Advance its Eliza AI Platform with Gemini Enterprise
- Agentic AI Arrives in Advisor Offices
- BNY gets access to OpenAI, Anthropic advanced cyber models
- Kinexys Brings Federated Learning for Fraud Detection
- Agents for financial services
- BNY Takes Aim At Schwab, Fidelity With AI-Powered Wealth Push
- AI in the C-suite: BNY CEO and OpenAI board member Robin Vince
- The Most Powerful Women in Banking Women to Watch, #18, Leigh-Ann Russell, BNY
- US 2025/0217603 A1: Large language model and neural networks for categorical classification of natural language text
- US 2025/0182530 A1: Adaptive, template-independent handwriting extraction from images using machine learning models
- US 12,417,410 B2: Generating optimal data predictions in real time for time series data signals
- US 2026/0057262 A1: Orchestrating interaction with an artificial intelligence application
- EP 4758506 A1: Translating a first coding language into a second coding language
When BNY 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