'Safety first' and data-lineage first: a Buffalo tech hub, Copilot for 16,000 staff, a second-line AI risk function — and a CEO who says the bank will keep humming 'a familiar refrain'.
How is M&T using AI?
M&T's AI strategy is the product of an eight-year technology transformation run by Michael Wisler, a Capital One veteran who became CIO in 2018, brought technology in-house, and built tech hubs in Buffalo and Wilmington that plan to hire 1,000 technologists within three to five years. The AI layer on top is deliberately careful: Microsoft Copilot for 16,000 of 22,000 employees, chosen for stability, on a data-lineage foundation that chief data officer Andrew Foster describes as 'safety first'; a dedicated second-line AI Risk Oversight function led by the former head of model risk; a head of AI engineering hired from Fannie Mae in August 2026; and agentic AI being examined for cybersecurity and fraud. CEO René Jones's 2026 letter sets the tone — 'while others chase the new siren song, we will continue to steadfastly hum a familiar refrain' — and the bank's award-winning credit-monitoring work with RDC.AI shows where AI is landing first: commercial portfolio insight.
Lead bank
Manufacturers and Traders Trust Company
Assets (lead bank, Mar 31, 2026)
$214 billion
Charter
State member bank (Fed)
Headquarters
Buffalo, NY
Flagship platform
Microsoft Copilot on a data-lineage foundation
AI leadership
Michael Wisler · Andrew Foster · Rhea Rajwani · Kalyana Bedhu · René Jones · Daryl Bible · Matt McAfee
- Eight-year transformation under CIO Michael Wisler (from 2018): technology brought in-house, contractor reliance cut, tech hubs at Seneca One in Buffalo (opened 2021) and in Wilmington, with a 2021 plan to hire more than 1,000 technologists over three to five years.
- Copilot for 16,000 of 22,000 employees — first drafts of emails and reports, call-centre summaries saving about six minutes per call; staff remain responsible for reviewing and approving outputs.
- Data before models: a data-lineage program, the Edison repository of authoritative documents, a Data Academy with about 2,000 graduates; Foster's 'safety first' framing.
- Governance built into the second line: Rhea Rajwani, formerly head of model risk management, leads AI Risk Oversight in independent risk management, aligned to regulatory expectations and peer benchmarks.
- August 2026: Kalyana Bedhu, from Fannie Mae's AI and machine-learning operations, named head of AI engineering — platforms, architecture and governance for AI across the bank.
- Where AI lands: credit monitoring and portfolio insight with RDC.AI (2026 Banking Tech Awards Editor's Choice); AI-powered cash application for commercial clients; participation in the Cari tokenised-deposit network and FIS's Project Keystone.
- Scale behind it: technology spending above $1.2 billion in 2025, about 2,000 technologists in more than 300 agile teams, and customer-affecting releases up from about 15,000 a year to 65,000; Jones says about 16,000 of 22,000 employees are now proficient in some form of AI, with LLM experiments under way in credit (September 2026).
- Disclosure: the 2025 10-K carries a dedicated AI risk factor (evolving law, incorrect or biased model output, limited explainability, dependence on third-party models, AI-enabled fraud); on the Q1 2026 call the CFO named automating operations 'using AI and other tools' as a priority for freed-up tech spend.
- Open AI roles on its careers site include a head of forward-deployed engineering for AI, an AI governance lead, a GenAI platform engineering lead and a model-risk validator for AI/ML and GenAI (listed below, as of 24 September 2026).
What is Microsoft Copilot on a data-lineage foundation?
M&T chose Microsoft Copilot as its enterprise generative-AI tool after chief data officer Andrew Foster vetted providers for 'a stable, strong partner'; 16,000 of 22,000 employees use it for first drafts and call summaries, which Foster estimates save about six minutes per call. Underneath sits a data-lineage program tracing where information originates and how it moves, an internal repository of authoritative documents called Edison, and a Data Academy that has trained about 2,000 employees.[1][4]
What has M&T done on AI, and when?
- Sep 18, 2026
Jones: 16,000 employees AI-proficient, LLM experiments in credit[17]
AI in cybersecurity, fraud and credit, anomaly detection to spot commercial-credit changes sooner, and a review of how vendors add AI.
- Sep 4, 2026
Three routes to generative AI[18]
General employee use, AI embedded in vendor applications and proprietary systems on the bank's own data; technology spend above $1.2 billion in 2025.
- Sep 2, 2026
Copilots to 15,000-plus employees; agentic AI examined for cyber and fraud[4]
Fast Company's Pacesetters profile of Wisler's transformation.
- Aug 19, 2026
RDC.AI credit-monitoring work wins Editor's Choice[8]
2026 Banking Tech Awards USA; earlier issue detection, fewer unnecessary alerts.
- Aug 18, 2026
Tokenised deposits: 'years of thoughtful work'[3]
Matt McAfee, head of enterprise innovation and digital assets, on managing the risks.
- Aug 11, 2026
Head of AI engineering hired from Fannie Mae[6]
Kalyana Bedhu takes engineering, platforms, architecture and governance for AI.
- Jul 23, 2026
AI Risk Oversight function profiled[2]
Rhea Rajwani builds the second-line AI risk program.
- Apr 30, 2026
Project Keystone: bank-owned tokenised money with FIS[12]
Citizens, Fifth Third, Huntington, KeyBank and M&T among participants.
- Apr 15, 2026
Q1 2026 call: freed-up tech spend goes to automation with AI[15]
General-ledger conversion complete; priorities are 'teaming for growth' and operational excellence using AI and other tools.
- Mar 16, 2026
Jones: AI aimed at the fundamental banking risks[16]
In an American Banker interview, AI investment focused on credit, liquidity, fraud and cyber rather than matching the largest banks.
- Mar 16, 2026
CEO letter: technology transformation, 'a familiar refrain'[7]
Jones on hiring and cultivating the best people to navigate change.
- Feb 18, 2026
Joins the Cari tokenised-deposit network[11]
With Huntington, First Horizon, KeyCorp and Old National.
- Feb 18, 2026
10-K adds a dedicated AI risk factor[14]
Evolving AI law, incorrect or biased output, limited explainability, reliance on third-party models and AI-enabled fraud and cyberattacks.
- Nov 6, 2025
Jones: no interest in national scale[13]
Focus on dominating current markets — $211 billion in assets, 960 branches, 13 states.
- Sep 18, 2025
Copilot for 16,000 employees; data quality first[1]
Foster vets LLM providers, chooses Microsoft; lineage and Edison underpin gen AI.
- Aug 5, 2025
Amperity chosen for a unified customer view[20]
The AI-powered customer data cloud consolidates customer interactions for M&T's Customer 360 work.
- Jul 16, 2025
AI and automation for working capital[10]
AI/ML cash application for commercial clients: remittance capture, invoice matching, deductions.
- May 15, 2024
Continuous credit monitoring on nCino with RDC's explainable AI[19]
Early-warning signs and cash-flow and credit-risk insight across commercial and business-banking portfolios.
- Apr 28, 2021
Buffalo Tech Hub opens at Seneca One[9]
Nearly $58 million invested in a 330,000-square-foot hub; plans to hire more than 1,000 technologists over three to five years.
- Jan 1, 2018
Michael Wisler becomes CIO[4]
Begins bringing technology in-house and reducing contractor reliance.
Where does M&T use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| Microsoft Copilot[1] | First drafts and call summaries for 16,000 employees. | In production | Generative & agentic AI |
| Credit monitoring with RDC.AI[8] | Agentic and predictive intelligence on commercial portfolios; earlier issue detection. | In production | Credit scoring & underwriting |
| AI cash application[10] | Remittance digitisation and invoice matching for commercial clients. | In production | Generative & agentic AI |
| AI Risk Oversight (second line)[2] | Independent oversight program for responsible AI adoption. | In production | Model risk management |
| Agentic AI for cyber and fraud (under review)[4] | Applications being examined, per Fast Company. | Pilot | Cybersecurity |
| AI anomaly detection and LLM experiments in credit[17] | Faster credit decisions and earlier warning of changes in a commercial customer's activity; experimental large-language-model work on credit applications. | Pilot | Credit scoring & underwriting |
| AI inside vendor applications[18][17] | One of three gen-AI routes across 1,800-plus applications, many vendor-supplied; the bank is reviewing how vendors incorporate AI. | Rolling out | Third-party & vendor AI |
| AI risk-factor disclosure[14] | 10-K covers evolving AI law, erroneous or biased output, explainability, third-party model dependence and AI-enabled fraud. | In production | Model risk management |
| Data lineage, Edison and the Data Academy[4][1] | The data foundation under the AI program; ~2,000 employees trained. | In production | AI governance (general) |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
Which AI vendors and fintechs does M&T work with?
| Company | What M&T uses it for | Status | Category |
|---|---|---|---|
| Pindrop | Call-center authentication and fraud detection; Pindrop describes M&T as a customer that switched from its existing authentication solution to Pindrop technologies. | In production | Fraud detection |
| Microsoft 365 Copilot | Microsoft Copilot chosen as the bank's generative-AI tool; 16,000 of 22,000 employees use it for first drafts and call-centre summaries. | In production | Enterprise AI assistants |
| nCino | Continuous Credit Monitoring powered by Rich Data Co's explainable AI (early-warning credit risk) | Announced | Banking platforms |
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 M&T disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Employees on Copilot | 16,000 of 22,000[1] | Sep 18, 2025 |
| Time saved per call (gen-AI summaries) | ≈ 6 minutes[4] | Sep 2, 2026 |
| Data Academy graduates | ≈ 2,000[4] | Sep 2, 2026 |
| Technologists to hire (Tech Hub plan) | 1,000+ over three to five years[9] | Apr 28, 2021 |
| Total assets | $213.8 billion[21] | May 6, 2026 |
| Employees proficient in some form of AI | ≈ 16,000 of 22,000[17] | Sep 18, 2026 |
| Technology spending | > $1.2 billion in 2025 (nearly 3x 2017)[18] | Dec 31, 2025 |
| Technologists / agile teams | ≈ 2,000 / 300+[18] | Sep 4, 2026 |
| Customer-affecting releases a year | ≈ 15,000 (five or six years ago) → 65,000[17] | Sep 18, 2026 |
“While others chase the new siren song, we will continue to steadfastly hum a familiar refrain.”[7]
René Jones · Chairman and CEO · Mar 16, 2026
“So our goal is to be sharper and reinvent and re-engineer those processes that are tied to those eight fundamental risks.”[16]
René Jones · Chairman and CEO · Mar 16, 2026
“Priorities now: teaming for growth—deeper wallet from customers and in regions—and operational excellence—simplify and automate operations using AI and other tools.”[15]
Daryl Bible · Senior EVP and CFO · Apr 15, 2026
“There's a lot of innovation here, but it's going to take years of thoughtful work, both from banks and the regulators that oversee us.”[3]
Matt McAfee · Head of Enterprise Innovation and Digital Assets · Aug 18, 2026
Who runs AI at M&T?
Michael WislerLinkedIn ↗
Senior EVP, Technology and Operations (CIO since 2018)[4][5]
Andrew FosterLinkedIn ↗
Chief Data Officer[1]
Rhea RajwaniLinkedIn ↗
Head of AI Risk Oversight, Independent Risk Management[2]
Kalyana BedhuLinkedIn ↗
Head of AI Engineering (from August 2026)[6]
Daryl Bible
Senior EVP and Chief Financial Officer[15]
Matt McAfeeLinkedIn ↗
Head of Enterprise Innovation and Digital Assets[3]
What AI roles is M&T hiring for?
8 open AI-related postings on M&T's careers site as of Oct 1, 2026. A snapshot of current openings, not a source: postings close.
Principal First Line Risk Specialist - Cybersecurity and AI Initiatives
Buffalo, NY · posted Sep 23, 2026
Model Risk Senior Analyst - Validation (AIML, GenAI, Technology)
Wilmington, DE · posted Sep 23, 2026
Lead Software Engineer (AI-Driven Development)
Buffalo, NY · posted Sep 29, 2026
Buffalo, NY
GenAI Platform Engineering Lead
Buffalo, NY
SDLC GenAI Automation & Tooling Integrations Engineer
Buffalo, NY · posted Sep 11, 2026
Data Scientist – Technology and Enterprise Operations First Line Risk
Buffalo, NY · posted Sep 14, 2026
Data Scientist – Technology and Enterprise Operations First Line Risk
Wilmington, DE · posted Sep 14, 2026
Which regulators govern M&T's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| Federal Reserve | State member bank supervised by the Fed; SR 26-2 and SR 23-4 govern models and the Microsoft dependency. | SR 26-2 SR 23-4 |
| NY DFS | New York-chartered: Part 500 and the DFS AI letters apply directly. | 23 NYCRR Part 500 DFS AI Cybersecurity Industry Letter (Oct 2024) |
| Basel Committee | The data-lineage program is BCBS 239 discipline applied to AI. | BCBS 239 |
| CFPB | Consumer and small-business lending; adverse-action duties apply. | ECOA / Regulation B adverse action (15 U.S.C. 1691(d); 12 CFR 1002.9) |
Side by side: SR 11-7 vs SR 26-2 · PRA SS1/23 vs SR 26-2 · OCC vs CFPB on AI lending · all comparisons.
What does the public record suggest about M&T's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
A second-line AI risk function is what supervisors describe, built before the agents
Moving the head of model risk into AI Risk Oversight is the organisational move the FSB's practices and the revised US guidance imply. Few regionals have done it this explicitly.
02
One vendor, chosen for stability, keeps the model risk simple
Copilot as the single enterprise tool trades capability for control. The third-party question is concentration in Microsoft; the model-risk question is easier than at multi-model peers.
03
The 'familiar refrain' is a strategy, not a lack of one
Jones's letter positions people and judgment over technology; the Buffalo hiring plan shows the bank means to own its technology anyway.
04
Watch the new AI engineering head
A Fannie Mae ML-operations veteran running platforms and governance suggests the next phase is in-house model building, not only Copilot adoption.
What AI does M&T Bank use?
M&T uses Microsoft Copilot as its enterprise generative-AI tool — 16,000 of its 22,000 employees use it for first drafts of emails and reports and for summarising call-centre conversations — on top of a data-lineage foundation, an internal document repository called Edison and a Data Academy. It also uses AI in commercial credit monitoring with RDC.AI and in cash application for commercial clients, and is examining agentic AI for cybersecurity and fraud.
How does M&T govern AI risk?
Through a dedicated second-line AI Risk Oversight function inside independent risk management, led by Rhea Rajwani, who previously ran the bank's model risk management. The program is aligned to regulatory expectations, industry guidance and peer benchmarks, and staff remain responsible for reviewing and approving AI outputs.
- How M&T Bank ensures data quality as it implements gen AI
- Rhea Rajwani — The Most Powerful Women in Banking (profile)
- Tokenized deposits are here. Banks need to manage the risks.
- How this Buffalo-based bank is using tech as a competitive edge
- How M&T Bank Built A Technology Powerhouse In Buffalo
- M&T Bank Appoints Fannie Mae Veteran Kalyana Bedhu as AI Engineering Head
- M&T Bank CEO René Jones Highlights Strong Performance, Technology Transformation in Annual Letter
- M&T Bank and RDC.AI Recognized with Editor's Choice Award at 2026 Banking Tech Awards USA
- M&T Bank Unveils Buffalo Tech Hub, Designed to Create Over 1,000 Jobs
- How AI and automation are unlocking working capital
- US Banks Build Tokenized Deposit Network to Guard Their Turf
- FIS and Leading Financial Institutions to Build Their Own Digital Tokenized Money Network
- M&T eschews the temptation of national presence
- M&T Bank Corporation Annual Report on Form 10-K for 2025
- M&T Bank (MTB) Q1 2026 Earnings Call Transcript
- Regional banks in today's economy: M&T CEO Rene Jones
- M&T Bank Sees Broad 2026 Loan Growth as AI and Fee Businesses Gain Traction
- M&T Bank expands enterprise AI after years of technology overhaul
- M&T Bank Expands Use of nCino with Adoption of Continuous Credit Monitoring Solution Powered by Rich Data Co's Explainable AI Platform
- M&T Bank Selects Amperity to Accelerate Customer Data Strategy
- M&T takes 'patient' approach to M&A, CFO says
When M&T 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