M&T Bank is aggressively integrating artificial intelligence into its operations, rolling out AI copilots to over 15,000 employees as it seeks to enhance efficiency and customer service across a broad spectrum of its business. This strategic deployment spans critical areas including internal operations, client interactions, software development, and risk management, signaling a significant shift towards an AI-augmented workforce.
The bank is leveraging AI for a variety of sophisticated tasks. Its call-center operations are being transformed through AI’s ability to analyze conversations, providing insights and identifying areas for improvement. Internally, AI assists in drafting reports and generating code, streamlining workflows for developers. Beyond operational efficiencies, AI is being employed to proactively identify customer needs and to flag potential risks within investment portfolios. M&T is also actively exploring the application of agentic AI in crucial cybersecurity and fraud detection initiatives, underscoring a commitment to robust security measures.
As reported in September 2025, M&T Bank had already made significant strides, with an impressive 16,000 of its approximately 22,000 employees actively using Microsoft Copilot. This widespread adoption focused on everyday tasks such as drafting emails, generating reports, and efficiently summarizing lengthy call-center transcripts, demonstrating a tangible impact on employee productivity.
Prior to this broad rollout, M&T Bank exercised caution by initially restricting employee access to public large language models. Andrew Foster, the bank’s Chief Data Officer, explained this measured approach was to mitigate the risk of sensitive company information being inadvertently exposed through public-facing AI services. This prudence allowed the bank to thoroughly evaluate enterprise-grade solutions. M&T subsequently selected Microsoft Copilot after a pilot program involving around 800 employees, a move that paved the way for its organization-wide deployment.
The benefits of this AI integration are already becoming evident. Foster highlighted that the use of generative AI to summarize call-center conversations can save approximately six minutes per call, a significant gain in operational efficiency. Similarly, software developers are utilizing GitLab tools to accelerate code generation, although M&T maintains a crucial human oversight layer, requiring employees to review all AI-generated code to ensure accuracy and adherence to standards. This emphasis on human review is further codified in the bank’s 2026 Code of Business Conduct and Ethics, which mandates the use of approved AI tools and strictly prohibits the input of confidential, proprietary, customer, employee, or regulated information into unapproved systems. Ultimately, employees retain full accountability for the accuracy and appropriateness of all AI-assisted work.
Building the Technology and Data Foundation
M&T Bank’s current AI initiatives are built upon a robust foundation established through a comprehensive technology overhaul that commenced in 2018. At the outset of this transformation, more than half of the bank’s technology specialists were external contractors. Today, this dynamic has shifted dramatically, with an overwhelming 80% of its technology workforce now being in-house. This strategic pivot has led to the creation of approximately 2,000 technologists organized into over 300 agile teams. The program has also seen the recruitment of more than 1,000 new technology specialists.
This modernization effort involved the replacement of numerous legacy platforms. The impact has been profound, with M&T reporting an over 80% reduction in technology outages since 2018. Concurrently, the frequency of system upgrades has surged by an impressive 300%. Technology spending has also seen a significant escalation, exceeding $1.2 billion in 2025, nearly tripling its 2017 expenditure. According to insights shared in August 2026, the bank’s annual technology releases dramatically increased from around 15,000 in 2018 to an impressive 65,000 by 2025.
The leadership driving this transformation is notable. M&T’s current Senior Executive Vice-President for Technology and Operations, who joined as Chief Information Officer in 2018, assumed broader responsibilities for both technology and operational functions in 2025, overseeing this critical integration.
The bank’s data program has been developed in parallel with its broader technology modernization. Since joining in 2023, the Chief Data Officer has been instrumental in establishing a data-lineage program. This initiative meticulously tracks the origin of data, its utilization, and its movement across various systems. Foster clarified that this data-lineage work predates the generative AI push and is considered a fundamental capability for understanding the bank’s extensive data landscape.
To foster data proficiency, M&T established a dedicated Data Academy, focusing on data governance and essential data skills. Approximately 2,000 employees have participated in this program, enhancing the bank’s collective data intelligence. Furthermore, M&T has developed “Edison,” an internal repository that serves as a centralized source for authoritative documents and information on bank policies. The bank also utilizes specialized data-lineage software from industry leaders Solidatus and Monte Carlo to map the flow of information through its databases, applications, and business intelligence systems. This detailed lineage mapping provides M&T with critical visibility into the source, meaning, quality, and governance of each data element. Foster noted that this governed data is foundational to the effective application of tools like Copilot. M&T is also employing retrieval-augmented generation techniques, utilizing its own internal, governed data sources.
Scaling AI into Daily Banking Operations
M&T Bank’s strategy for adopting generative AI is multifaceted, encompassing three primary avenues: general employee utilization, integration of AI capabilities into existing applications, and the development of proprietary systems tailored to the bank’s unique data and processes.
With an operational portfolio exceeding 1,800 applications, many sourced from third-party vendors, one of M&T’s AI strategies involves identifying and leveraging AI functionalities already embedded within these existing platforms.
The third prong of M&T’s approach centers on proprietary AI development, focusing on solutions built around the bank’s proprietary data and operational workflows. Initial applications in this domain are reportedly targeting repetitive operational tasks, accelerating software development cycles, enhancing fraud prevention mechanisms, and bolstering cyber defenses. M&T continues to evaluate a spectrum of AI solutions, including internally developed systems, general enterprise software, and specialized financial technology offerings.
Early AI use cases have primarily focused on content generation, summarization, call-center support, and software development. More recent advancements have expanded to identifying customer needs and proactively flagging portfolio risks.
M&T’s proactive AI integration mirrors a broader trend among major U.S. financial institutions. JPMorgan Chase, for instance, launched its internal LLM Suite platform in 2024, making it available to over 200,000 employees. By 2025, its Corporate and Investment Bank saw active engagement from more than 65,000 employees, with over 90% of its engineers utilizing AI coding assistants. The bank also reported that its AI-powered transaction screening capabilities have doubled the review volume of transactions while halving the reliance on manual operator checks.
Bank of America has deployed EricaAssist, a generative AI-enabled system, to more than 18,000 customer service employees. This tool is designed to provide immediate context by summarizing customer inquiries, retrieving relevant information, and recommending next steps, while maintaining human accountability for the overall customer interaction. Bank of America indicated in July 2026 that EricaAssist can deliver contextual guidance in under three seconds, resulting in an average reduction of nearly one minute per call. The bank plans to expand the system’s reach to additional servicing scenarios and business lines later in 2026, further solidifying AI’s role in enhancing customer engagement and operational efficiency.
Original article, Author: Samuel Thompson. If you wish to reprint this article, please indicate the source:https://aicnbc.com/25465.html
