BNY Leverages Azure OpenAI for Financial Institution AI

BNY, America's oldest bank, is scaling AI with Microsoft 365 Copilot and AI agents across the business, supported by an AI bootcamp that has trained more than 2,300 employees. The result: stronger governance, faster decisions, and more time for people to focus on high-value work. If a bank founded in 1784 can become a Frontier Firm, what's stopping the rest of the industry?

If you knew that what you were building could carry enormous risks, would you choose to slow down? I’ve been thinking about this after seeing some of the recent discussion around AI. The leaders of Anthropic and OpenAI have started publicly discussing whether the development of the most advanced AI models should be slowed down in the interest of safety. It sounds reasonable in theory, but putting it into practice is a lot harder. AI is no longer just a technology issue. There are investors, governments, chipmakers, cloud providers, and an entire industry competing to move faster. What happens if you slow down, but your competitors don’t? I think that’s what makes this debate so difficult. The real challenge isn’t that people don’t see the risks of AI. It’s that even when everyone recognizes the risks, it’s incredibly hard to be the one who actually hits the brakes. When everyone is racing forward, the first person to slow down may also be the one who pays the biggest price. What do you think? If you were running an AI company, would you keep pushing forward at full speed, or would you be willing to slow down a littl

This is the room where it happens. Nearly 250 years after Alexander Hamilton founded it, America's oldest bank is using AI and superagents to absorb the work that used to grind at attention and patience—and ultimately make work more human. Go behind the scenes of BNY big bet on AI with our reporter, who spent two days at the bank's Manhattan HQ. Learn more: https://msft.it/6043vRGuV

Giving employees access to AI is simple. The real challenge is creating an environment where they can learn, experiment, and use it responsibly. Using Microsoft AI, BNY connects training, development, and governance into a single operating model that helps employees turn AI skills into business outcomes. Agents built through this approach already triage stalled payments and resolve client inquiries in production. Access opens the door. Connecting learning, building, and governance is what turns AI adoption into real business impact.

Everyone keeps talking about making AI models bigger. I’m becoming more interested in making them smaller. I’ve been playing around more with local AI recently, and one thing that surprised me is how much you can actually run on a normal laptop. The trick? Quantization. In simple terms, you take a model that might normally need a ridiculous amount of memory and compress it so it can run with much less. And the interesting part is that the model doesn’t suddenly become useless. You’re trading a little bit of quality for a lot less memory. Think of it like compressing a huge image. You might lose some detail, but if you’re looking at it on your phone, you might not even notice. This makes local AI much more interesting for businesses. You can have models running: → on your own hardware → without sending data to an external API → without paying per API call → with much more control over your data And you don’t necessarily need a €10,000 GPU server to do it. For a lot of internal AI use cases, you probably don’t need the biggest model available anyway. You need a model that is: Fast enough. Good enough. Cheap enough. Private enough. That’s a very different optimization problem. I think we’re going to see a lot more AI moving from “send everything to the biggest model in the cloud” toward small, specialized models running closer to where the data actually lives. The future of AI might not just be bigger models. It might be more flexible models, running everywhere.

How can AI help transform operations at one of the world's largest financial institutions? BNY is combining its proprietary AI platform, Eliza, with Microsoft AI technologies to reimagine how work gets done across the organization. Built on Microsoft Azure and integrated with tools such as Microsoft 365 Copilot, GitHub Copilot, and Microsoft Foundry, Eliza helps employees innovate, automate complex processes, and deliver faster results for clients. In this customer story, explore how BNY is applying AI across the financial services value chain, from client onboarding and transaction support to software development and employee experiences. By embedding AI into everyday workflows, BNY is helping teams work more efficiently, improve service delivery, and unlock new opportunities for innovation. #MicrosoftAdvocate #MicrosoftEmployee

We’re seeing it: AI is transforming even the most regulated industries. A 240-year-old bank like BNY is leading—using AI agents to handle repetitive work and empower people to focus on strategy and decision-making. It’s not about replacing talent, but elevating it. Curious what this means for your organization? Read the latest Work Trend Index.👇 https://msft.it/6041ayNXZ #WeAreMicrosoft #MicrosoftAdvocate

🚀 AI-first “Frontier Firm”: a look at how #BNY is redesigning work with #AI—as a transformation engine: 🧠 AI is for everyone: Every employee is trained to design work, not just do it 🤖 Digital employees: 140+ AI agents operate like team members with roles, managers, KPIs ⚡ Massive efficiency gains: Tasks go from mins → secs 🔄 Shift from repetitive → strategic work: humans move up value chain (analysis, strategy, design) 🧩 New org model: from pyramid → diamond (less repetition, more creativity and judgment) 🎓 AI upskilling at scale: thousands trained, with real prototypes built by employees 🔐 Trust-first AI with governance, compliance, and observability built into every step ⚙️ AI building AI: workflows can now be captured and turned into automated agents in hours 💡 Takeaway: This isn’t about doing less work — it’s expanding what humans are capable of doing. Learn more: https://msft.it/6040vRC2w #FutureOfWork #DigitalTransformation #Copilot #Leadership #Innovation #FinancialServices #AgenticAI #WorkplaceEvolution #FrontierFirms #MicrosoftAI