ConnectOne Bank is bringing AI agents directly into its commercial lending operations.
Not as a separate experiment. Not as another chatbot sitting in the corner of the bank’s website.
The New Jersey-based lender is expanding its long-running partnership with banking software provider nCino, deploying two AI agents built on the company’s Agentic Operating System. The tools are designed to handle some of the repetitive, slow and frustrating work that still fills a commercial lender’s day.
That could mean less time digging through files, repeating administrative steps or tracking down information buried somewhere inside a lending platform.
It sounds modest. Inside a bank, it is not.
ConnectOne Brings AI Agents Into Daily Lending Work
The two AI agents currently include 16 specialised skills focused on commercial lending workflows.
nCino says the agents are intended to automate high-friction and time-consuming tasks across the lending process. More capabilities are expected to be added over the coming months as ConnectOne expands how the technology is used.
The announcement does not provide a complete list of every task the agents will perform. Still, the direction is fairly clear.
Commercial lending involves a lot of document gathering, internal searches, portfolio reviews, application checks and follow-up work. Much of it is necessary. Very little of it is exciting.
ConnectOne appears to be targeting that operational layer first.
Rather than asking AI to make every lending decision, the bank is using it to clear away work that slows employees down before a decision can even be made.
Document Searches Have Reportedly Dropped From 20 Minutes to 30 Seconds
One number stands out.
ConnectOne has also integrated nCino’s Banking Advisor, a generative AI-powered conversational assistant, together with the company’s Knowledge Base capability.
According to nCino, those tools have reduced the time employees spend searching for documents by 97.5%. A lookup that previously took around 20 minutes can now reportedly be completed in roughly 30 seconds.
Vendor-reported efficiency figures always deserve some caution. Even so, the size of the claimed reduction is difficult to ignore.
A single 20-minute document search may not sound disastrous. Multiply it across lending teams, active borrowers, credit reviews and an entire working year. Suddenly it becomes a serious operational cost.
Thirty seconds changes the rhythm of the work.
Employees can spend less time hunting for information and more time reviewing the borrower, checking risks or speaking with clients. At least, that is the practical promise.
The nCino Partnership Goes Back to 2017
ConnectOne and nCino are not new partners.
The bank first adopted nCino’s Bank Operating System in 2017 to support account opening, lending and portfolio management. The latest AI rollout builds on nearly a decade of shared technology infrastructure rather than replacing the bank’s systems overnight.
That matters.
AI projects inside banks are much harder when the technology has to be connected to fragmented data, old software and unfamiliar workflows. ConnectOne already has nCino embedded across parts of its operation, giving the new agents a more realistic path into everyday use.
ConnectOne chairman and CEO Frank Sorrentino III said the original implementation was meant to preserve the bank’s efficient operating model while supporting growth and faster delivery to the market.
The current phase goes further. The bank is now looking at how AI might reshape the way its teams work, not simply digitise the processes they already have.
Commercial Lending Is Becoming a Serious Test for Agentic AI
Commercial lending is a useful place to test AI agents because the work is complicated but full of repeatable steps.
Applications arrive with financial statements, legal documents, ownership records, collateral details, credit reports and internal notes. Information has to move between relationship managers, analysts, underwriters, operations teams and compliance staff.
Plenty of room for delays.
AI agents could help organise that movement, retrieve the right information and complete structured tasks without requiring an employee to manually guide every step.
The difficult part will be control.
Banks cannot treat lending automation like ordinary office software. Decisions need to be explainable. Customer information needs protection. Employees must know when an AI-generated response is reliable and when it needs a second look.
ConnectOne’s initial focus on administrative friction may be the sensible route. Start with the tedious work. Measure it. Keep humans close to the credit decisions.
Then expand carefully.
ConnectOne Has Grown Into a $14 Billion Bank
Founded in 2005, ConnectOne operates 57 branches across New York, New Jersey and Long Island. The institution manages approximately $14 billion in assets and provides services including commercial loans, deposit accounts, residential mortgages, online banking and payment processing.
This is not one of the largest banking groups in the United States.
That may be exactly why the deployment matters.
Regional and mid-sized banks are under pressure to deliver the kind of fast, digital experience customers now expect without carrying the technology budgets of the biggest global institutions.
AI agents may give those lenders another way to compete. Not by becoming fully automated banks, but by reducing the amount of manual work sitting between a customer request and an employee’s response.
AI in Banking Is Moving Past the Chatbot Stage
For years, banking AI was mostly discussed through customer-service chatbots, fraud alerts and personalised product recommendations.
The industry is moving into a more operational phase.
AI tools are entering document management, underwriting preparation, compliance reviews, internal knowledge systems and lending workflows. nCino’s recent work with other financial institutions also shows that AI-supported commercial lending is becoming part of broader banking infrastructure modernisation rather than a stand-alone novelty.
ConnectOne’s deployment fits that shift.
The agents are not being introduced simply to answer questions. They are being given skills, access to lending workflows and specific jobs to complete.
That is where agentic AI starts to become interesting for banking.
It is also where the risks become more real.
What Comes Next for ConnectOne and nCino
The current rollout includes two AI agents and 16 specialised skills, but both companies have indicated that additional capabilities are planned.
The real test will not be how impressive the technology sounds during launch week.
It will be whether lending teams actually use it, whether the agents produce consistent results and whether the reported time savings continue once the system is operating across more complicated cases.
Commercial loans are rarely simple. Borrowers are different. Documents are messy. Exceptions appear everywhere.
AI agents will have to work inside that reality.
Still, cutting a document search from 20 minutes to around 30 seconds is a strong place to begin.
For ConnectOne, the goal appears straightforward: remove the operational drag without losing the judgment that commercial banking still requires.
That balance will decide whether the rollout becomes a useful banking upgrade or merely another AI announcement.
Sources
- FinTech Futures – ConnectOne rolls out nCino AI agents across commercial lending
- FinTech Futures – Norway’s DNB picks nCino for corporate lending revamp