While most people perceive artificial intelligence through interfaces like ChatGPT or Claude, the world of big finance is looking in a completely different direction. It's not about AI writing an email, but about it being able to independently process a loan application, verify a client's creditworthiness, and prepare the final documentation. This is exactly what the platform from nCino, which ConnectOne Bank has started using, does.
What Is "Agentic AI" and Why Is It Different from ChatGPT?
To understand this shift, it's necessary to distinguish between generative AI and agentic AI (Agentic AI). Generative models, such as GPT-4 from OpenAI or Gemini from Google, are masters of text prediction and information processing. They are the "brain" waiting for instructions.
Agentic AI, however, is equipped with the ability to act. An agent is not just a model; it is a system that has access to tools (e.g., databases, calculators, email systems) and can plan steps to achieve a goal. If you tell a generative AI: "Write me a loan analysis," it will write you a text. If you tell the nCino agentic system: "Process this commercial loan application," it will log into systems itself, download financial statements, perform risk analysis, and prepare an approval proposal.
Compared to regular LLMs (Large Language Models), agentic systems in banking offer a higher degree of autonomy and workflow integration. While Claude 3.5 Sonnet excels in programming and logical reasoning, nCino uses these abilities specifically for orchestrating complex banking processes.
The ConnectOne Bank Case: Numbers That Make Sense
The implementation of the nCino Agentic Operating System at ConnectOne Bank is not just a theoretical experiment. According to reports from Yahoo Finance, this technology delivered immediate results in commercial banking:
- Productivity increase: Commercial bankers are now 50 % more efficient at portfolio management and processing new cases.
- Time savings: The time required to complete specific lending tasks has been reduced by 60 %.
- Process optimization: AI agents eliminate administrative "bottlenecks" that previously required manual data entry from various sources.
This effect is due to bankers no longer having to spend hours copying data between different systems. Agentic AI performs these steps in the background, while the human can focus on decision-making and building relationships with clients.
Technology Context: nCino and the Salesforce Ecosystem
It's important to mention that nCino is not built from scratch in isolation. The platform runs on the cloud foundation of Salesforce, the world leader in CRM (Customer Relationship Management). Thanks to this integration, the agentic AI has access to rich client data, which is critical for analysis accuracy. For comparison, while specialized fintech tools such as Abrigo or FIS offer strong analytical modules, nCino, with its agentic layer, aims to create a truly autonomous operating system for a bank.
Pricing and Availability
Tools like nCino are not intended for individuals or small business owners. This is an Enterprise SaaS (Software as a Service) solution with a very high entry barrier. Prices are not publicly available, but in this segment they range from tens to hundreds of thousands of dollars per year for a license, with pricing always individually negotiated based on the size of the bank and the scope of implementation. The tool is therefore not directly available to the average user.
Impact on the Czech and European Market
Although ConnectOne Bank operates in the USA, this development has two major implications for Czech and European banking institutions (such as Česká spořitelna, ČSOB, or Komerční banka):
- Regulation and the EU AI Act: In the European Union, the EU AI Act is coming into force, which classifies systems used in finance (especially for loan scoring) as high-risk. The implementation of agentic AI must therefore meet extreme demands for transparency and explainability. Banks in the Czech Republic will not be able to use a "black box," but must know why the agent decided the way it did.
- Competitive pressure: The Czech banking sector is highly digitized. If American and global banks shorten the loan approval process from weeks to hours thanks to agentic AI, European banks will have to follow suit in adopting similar technologies to remain competitive.
In the Czech environment, we do not yet see direct localization of nCino for the Czech language, but the technological trend is clear: the future of banking will not be about applications you click on, but about agents that work for you.
What about data security when using agentic AI in a bank?
Agentic systems like nCino run in closed cloud environments (e.g., Salesforce) that meet strict banking security and compliance standards. Data is not shared with public models like ChatGPT, but remains within the institution's controlled environment.
Can agentic AI replace bankers?
No, the goal is not to replace humans, but to eliminate routine, administrative tasks. Agentic AI acts as a "super-assistant" that prepares materials so the banker can make the final, responsible decision based on expertise and the client relationship.
Is agentic AI compliant with EU regulations?
The development of tools like nCino must adapt to the EU AI Act rules. This means systems must be designed so that their decision-making processes are auditable and transparent, which is a key requirement for the banking sector in Europe.
