nCino

Loan origination, account opening and portfolio monitoring are workflow software that runs without any model, and that workflow is what more than 2,700 institutions bought. The AI sits on top of it. Banking Advisor answers questions over loan files and credit policy, Digital Partners carry out tasks, and machine learning speeds spreading and credit monitoring. Switched off, the agents would leave the lending system of record intact.

The mortgage suite spells out its controls. Agents connected through Mortgage MCP act only within the user's existing permissions, and every action is logged with a timestamp, the action and its outcome. Institutions can require a person to confirm high impact actions, such as archiving a loan officer or restructuring a branch, before they run. Mortgage Advisor for Loan Officers creates tasks or sends reminders only with the loan officer's approval.

Across the wider platform, human review is the default, every Digital Partner action leaves a record, and the agents surface the points where a banker has to decide. Automated preapproval, which chains credit pulls, verifications, fees and two automated underwriting systems, is the mortgage path where nCino describes less originator involvement rather than a confirmation step.

nCino says it keeps model risk documentation aligned with guidance from the Office of the Comptroller of the Currency and the Consumer Financial Protection Bureau. That documentation is not public, and no validation result is published. The visible mechanisms sit in the product. Automated Spreading reconciles extracted figures line by line, flags fields it cannot read clearly and learns from earlier mappings.

Locate and File places documents in the loan file and extracts their data at 99 percent accuracy by nCino's own figure. A loan officer at Educated Mortgage reports that Mortgage Advisor cited the guideline behind its answer.

ConnectOne Bank runs Banking Advisor's knowledge base in production alongside two custom agents with 16 skills built on the Agentic Operating System. Searching credit policy and job aids fell from 20 minutes to as little as 30 seconds, active users grew 41 percent in ten weeks, and an agent that updates individual and business relationship records cut that task by 60 percent.

Frank Sorrentino III, the bank's chairman and chief executive, has set a goal of making every frontline banker 50 percent more productive. Educated Mortgage already uses the loan officer tools in daily origination. The speed figures nCino gives for Banking Advisor and the analyst agent come from internal, simulated testing, as nCino itself notes.

The line nCino draws is around personal data. It does not use personally identifiable or nonpublic personal information to train its agents or foundation models, and no institution's data trains the Banking Advisor language models. Pooled learning comes from elsewhere. More than 1,500 customers contribute data context to the platform, and institutions that join the nCino Data Community share anonymized operational data in exchange for a free tier of Operations Analytics.

Retention periods for client data are not public. Within those limits, nCino says its banking AI is grounded in more than 14 years of credit decisions, portfolio outcomes and workflow patterns rather than the open web.

Client institutions are the controllers and nCino the processor for the personal data it handles on their behalf, and it processes that data only on their instructions. Neither its data processing terms nor a subprocessor list is public, though nCino binds its subprocessors to written privacy, data protection and security obligations and audits them. Its US entities take part in the EU, UK and Swiss data privacy frameworks run with the US Department of Commerce. On the Gramm Leach Bliley Act (GLBA), nCino's position is that it never trains its models on nonpublic personal information, which keeps an institution's privacy disclosure straightforward.

nCino names an ISO 27001 certification, SOC 1 Type II and SOC 2 Type II reports, alignment with the Cloud Security Alliance Security, Trust, Assurance and Risk (STAR) program, and controls that meet the Payment Card Industry Data Security Standard for funding applications. The ISO 27001 certification is cited against the 2013 edition of the standard, and none of the reports or certificates is offered for public download. Day to day controls include least privilege access with single sign on, intrusion detection with continuous monitoring, layered infrastructure defenses and formal incident notification to clients.

As a technology provider rather than a lender, nCino says federal and state banking regulators, which it does not name, examine it periodically, including the information technology controls behind the data processing it performs for client institutions. That examination covers nCino as a service provider, not its AI features.

The platform ships compliance frameworks for integrated mortgage disclosures, Home Mortgage Disclosure Act (HMDA) and Community Reinvestment Act reporting, the small business lending data rule under section 1071 of Dodd Frank, know your customer and anti money laundering checks, and the EU Digital Operational Resilience Act.

Two of the agents work close to the credit decision. Loan PreCheck recommends a mortgage program from a borrower's eligibility, and Analyst Digital Partner speeds risk assessment and credit analysis for underwriters. nCino's own governance language covers audit trails, explainability and rule anchored logic.

Fair lending shows up in its material as reporting tools sold to client institutions, HMDA data and portfolio dashboards built for fair lending review, rather than as testing of its own models for disparate impact or adverse action reasons.

When an agent or a model gets something wrong, the institution carries the outcome. nCino publishes no guarantee, indemnity or accuracy commitment for its AI, and a borrower affected by a recommendation has no route of their own to nCino. What it supplies is a record to challenge from. Each Digital Partner action is written to a record, decisions can be traced end to end, and Mortgage MCP logs each action with its time and outcome in nCino's system.

Which model powers Banking Advisor, Mortgage Advisor or the Digital Partners is not stated. The infrastructure and some AI partners are named. Fundamental elements of the platform sit on the Salesforce Platform, AWS provides the cloud infrastructure for the intelligence capabilities, and Anthropic and Databricks appear among the technology partners nCino lists as integrated into the platform, alongside Alloy, Plaid and Docusign.

At most client institutions nCino is the lending system of record, so integration runs outward from it. Integration Gateway, formerly Glyue by Sandbox Banking, is an integration platform with prebuilt connectors for more than 14 core platforms, Fiserv, Jack Henry and FIS among them, and for more than 50 specialist partners, carrying both real time calls and batch transfers.

A developer portal opens the platform's interfaces to client engineers, and MCP servers let outside agents work inside the mortgage suite. Alloy, Plaid, Docusign and Rich Data Co are among the partners nCino lists as integrated.

The platform runs as a cloud service, and where client data rests is not stated for any market, though nCino sells in the US, the UK and Ireland, continental Europe, Japan and Australia. Salesforce supplies the hosting and data centers for the applications nCino builds on the Salesforce Platform, and AWS runs the data layer behind the intelligence features. Inside the service, each component is isolated by tenant identifier with its own encryption keys, and separate environments hold different functions.

Figures come only through a quote, but the billing units are public. Since fiscal 2025 nCino has moved customers to asset based pricing, so a subscription grows with the institution's portfolio assets and its use of AI capabilities. Digital Partners are metered in Intelligence Units under what nCino calls its Intelligence Solution Framework, and the total depends on which products an institution deploys. Operations Analytics has two tiers: a base tier free to institutions that join the nCino Data Community, and a Pro tier charged as a platform fee.

Community banks, credit unions, enterprise banks, independent mortgage banks, building societies and mutuals each get their own material, and nCino reports customers in 20 countries. Named users run from Wells Fargo, Bank of America, Barclays, Santander and TD Bank through Truist and U.S. Bank to WaFd Bank, Eastern Bank and Navy Federal Credit Union. Hiroshima Bank and Hachijuni Nagano Bank, two Japanese regional banks, chose it for consumer lending in 2026.

Product lines run from commercial, small business, consumer and mortgage lending to deposit account opening and portfolio analytics, with separate mortgage offers for North America, the UK and Ireland, and Australia and New Zealand.