🔒AI adoption accelerates across CT’s banking industry

When Ted Balagtas joined Ion Bank as its chief information officer in April 2025, artificial intelligence was already emerging as a strategic priority. At the time, the Naugatuck-headquartered bank was forming an internal committee to evaluate how AI tools could be introduced, while also developing policy and prioritizing data security — a balancing act confronting many community institutions. Ion, with $2.67 billion in assets, has begun experimenting with generative artificial intelligence, allowing employees to use Microsoft’s Copilot to draft communications and presentations and summarize documents, Balagtas said. The bank is now preparing to broaden that use, beginning with expanded Copilot training delivered through in-person sessions and webinars before moving into individual business lines to identify potential use cases, Balagtas said. “One of my primary goals for 2026 is identifying use cases and seeing what makes sense for our business models and workflows,” he said. Ion’s experience reflects a broader shift underway across Connecticut’s banking sector, where artificial intelligence is rapidly moving from experimentation to operational reality. For bank executives, the discussion is no longer whether AI tools will influence the industry, but how quickly they can be deployed without introducing regulatory, security or reputational risks. Bankers and industry analysts say that reality is intensifying competitive pressure, as institutions that fail to integrate AI tools risk ceding efficiency, speed and analytical advantages to rivals moving more aggressively. “You cannot resist the march of technology in something like this,” said John Carusone, president of the Bank Analysis Center in Hartford. “It's sort of like resisting the march of computers 50 years ago.”

Fintech partnerships

That transition, however, is unfolding unevenly across institutions, Carusone said. Large national banks are investing aggressively in artificial intelligence, while smaller and midsize institutions face tighter technology and staffing constraints. As a result, many community and regional banks are turning to fintech firms and technology vendors to accelerate AI adoption across their operations, including underwriting and investment management. Over the past two years, New York-based startup Kobalt Labs has built a client base of 50 banks, credit unions and fintech firms, offering AI-enabled products that automate risk and compliance reviews. It entered the Connecticut market late last year, signing New Canaan-based Bankwell Bank as a client. “Almost every banker has used AI, maybe not the entire bank has bought an enterprise license, but there are very few bankers left who have not dabbled with AI in some way,” said Kalyani Ramadurgam, co-founder and CEO of Kobalt Labs. “So, this is no longer a foreign concept.” Banks have traditionally relied on manual documentation reviews to vet vendors, conduct audits and ensure regulatory compliance. AI platforms now automate much of that labor-intensive work. “Some of our users today say that the nature of their job has completely transformed,” Ramadurgam said. “They used to spend time going control by control, reading through documents and checking boxes. Now, they don’t spend any time doing that anymore, and their job has shifted to looking at what AI has surfaced and then using their experience and industry knowledge to act on those key findings so they can get things done more quickly.”

Ryan Hildebrand is the chief innovation officer of Bankwell, which has $3.2 billion in assets. He said staff across age groups are eager to experiment with AI. His bank launched a pilot of Microsoft Copilot last year, beginning with 20 employees in marketing and credit operations. “And really that pilot group kind of exploded,” Hildebrand said. “Everyone wanted to get in, and it's become really a working group. We have about 175 employees, and I'd say a good portion are using Copilot every single day. And, for us, it's been really a cultural shift. We’re trying to encourage the mindset of everyone asking, ‘can AI help do this better or faster?’”

Testing and learning

That pattern — cautious experimentation followed by wider internal adoption — is playing out at institutions across Connecticut. At Norwich-based Dime Bank, executives are preparing to formalize that shift through board-level policy. President and CEO Nicholas J. Statoulas said the $1.2 billion institution expects to present a new AI policy to its board later this month designed to establish guardrails governing how employees use emerging AI tools while protecting sensitive customer information. Once approved, some staff will begin experimenting later this year with Microsoft’s Copilot AI assistant. Importantly, Copilot can be used to analyze internal Dime data securely without exporting information beyond the bank’s systems, Statoulas said. At Middletown-based Liberty Bank, which has $8.68 billion in assets, executives have formed an AI working group and launched a limited Copilot program. “It’s on a limited basis, but we’re testing and we’re learning how we can use that effectively in our environment,” said Patrick Torney, Liberty’s chief information officer. Torney noted that AI upgrades are increasingly embedded into platforms banks already use, including customer relationship management systems such as Salesforce, requiring institutions to maintain close oversight of vendor relationships.

Larger banks invest at scale

At larger institutions, the pace of AI deployment is typically faster. Ohio-based KeyBank, which holds $184.69 billion in assets and operates 47 branches in Connecticut, has been training staff in basic AI skills for three years. All employees gained access to Copilot last year after completing ethics and compliance training.

“We have built an enablement roadmap that tries to meet our teammates where they are, bootstrap their knowledge of Copilot through prompt libraries, team training sessions and ultimately (have) them find value from integrating Copilot into their daily work,” said Kim Snipes, head of commercial digital, AI and client experience for KeyBank. KeyBank is now targeting higher-return applications powered by agentic AI — autonomous systems designed to pursue specific goals with limited oversight. Snipes cited opportunities in software engineering, front-line sales, fraud and anti-money laundering prevention, servicing, operations and collections. New York-based NBT Bank — a $16 billion institution with five Connecticut branches — has used robotic process automation for two decades and is now enthusiastically embracing generative AI tools such as ChatGPT and Copilot, said NBT President Joseph Stagliano.

NBT is using tools like AI to automate workflows and generate business insights, freeing staff to focus on customer-facing activities and high-impact tasks, Stagliano said. At the bank’s annual leadership conference at the Albany Capital Center, Stagliano co-led a presentation titled “AI is for everyone.” To support adoption, NBT established an AI Think Tank comprising 50 leaders across the organization. The group shares use cases, conducts training sessions and fosters awareness of AI applications, both embedded in existing systems and available as standalone tools. Training is delivered through a mix of in-person workshops and online learning platforms. Stagliano said centralized data infrastructure and strong governance have enabled swift adoption.