Source: GlobeNewswire — 2026-07-14
Summary
ConnectOne Bank deployed nCino's Agentic Operating System across its commercial lending operations, combining a generative-AI copilot ("Banking Advisor") with two custom-built agents. The bank reports document lookup time dropped roughly 97.5% — from about 20 minutes to 30 seconds — while a separate Document Intelligence agent that updates individual and business relationship records cut task time by 60%, with the whole rollout built and refined in weeks rather than the usual multi-year transformation timeline.
Key Takeaways
- The core metric: document lookup time fell from ~20 minutes to ~30 seconds (a 97.5% reduction) using nCino's generative-AI copilot plus its Knowledge Base capability layered on top of the bank's existing document stores.
- A second, separately measured agent — built specifically to use "Document Intelligence" to update individual and business relationship records — cut task time by 60%, a distinct workflow from the lookup-time improvement above.
- nCino's forward-deployed engineering team worked directly inside ConnectOne's own environment to design and refine the agents, rather than shipping a generic off-the-shelf tool — the deployment model itself (embedded engineers iterating in production) is as notable as the resulting metrics.
- The bank frames this as compressing what's traditionally a multi-year core-banking transformation into an active, iterative rollout measured in weeks — a claim worth watching for follow-up data on durability and error rates as usage scales.
Reel Script
Hook (16s / 36 words) A commercial bank just cut a 20-minute manual task down to 30 seconds — not with a chatbot bolted onto the front end, but by giving loan officers an agent that actually knows where the documents live.
Core Concept (55s / 120 words) Commercial lending runs on documents — financial statements, entity records, prior loan history — scattered across systems that loan officers have to manually search every time they need to verify something. ConnectOne Bank built its fix on top of nCino's "Agentic Operating System," a platform specifically for banks to build and run AI agents against their existing loan and customer data, rather than a generic AI assistant. The key piece is a "Knowledge Base" layer feeding a conversational copilot called Banking Advisor: instead of a banker manually digging through folders and systems, they ask a question in plain language, and the agent retrieves the specific document or fact directly.
Hands-On (80s / 180 words) Two separate numbers came out of this deployment, and they're measuring two different things — worth keeping straight. First: document lookup time, the time it takes a banker to find a specific document they need, dropped from about 20 minutes down to roughly 30 seconds — a 97.5% cut. That's the Banking Advisor copilot plus Knowledge Base doing retrieval work a human used to do by hand. Second, and separately: a custom-built agent using "Document Intelligence" — reading incoming documents and automatically updating individual and business relationship records in the bank's systems — cut the task time for that specific update workflow by 60%. That's not retrieval, that's structured data entry being automated. What's also worth noting is the deployment model: nCino didn't just sell ConnectOne software — its own engineers worked embedded inside the bank's environment, building and refining these two agents directly against real production data, compressing a process banks usually measure in years into a rollout measured in weeks.
Takeaway (23s / 51 words) This is a concrete before-and-after in a regulated, document-heavy industry — not a hypothetical. If your team is evaluating whether agentic tooling is worth the integration cost in a similarly document-bound workflow, this is the kind of ROI case to benchmark against.