What did the bank launch?
According to the bank, the platform offers basic productivity scenarios and typical AI agentsAI agent. By definition, an AI agent is a neural network that does not just answer but carries out a task step by step on its own: it looks for data, uses programs and checks the result. for different office roles out of the box. A company can manage the agents that employees create, set rules for their work, access control and moderationAI content moderation. Content moderation is the check of messages, reviews, images and AI answers for banned and harmful content, before or after publication. AI content moderation does this check with a program.. Content moderation is a check of messages, reviews, images and AI answers for banned or harmful content before or after publication. The platform was built by the bank group’s business-technology company.
A senior vice president of the bank, who leads its technology development, said the platform lets a company take into account its regulations, knowledge baseAI knowledge base. An AI knowledge base is a set of selected and prepared company documents that an AI assistant searches for answers: rules, instructions, a product catalog and answers to common questions., client request system and other features of its processes.
What are the three formats and the promised effect?
- Cloud (SaaS) for those who need fast scaling.
- Hybrid in a private cloud: data stays on the company’s servers, and the language modellarge language model. A large language model (LLM) is a neural network trained on a huge amount of text. It writes, summarizes and analyzes text on request. does not keep information after it answers. A large language model is a neural network trained on huge amounts of text that writes, summarizes and analyzes text on request.
- On-premises hardware and software complex for large companies, banks and critical information infrastructure.
The bank says configurable AI agents inside the company perimeter can cut staff time spent on routine tasks by up to 70%. This is the developer’s estimate, not a measurement on your process.
What does it mean for a small or medium business?
A large bank has made AI agents a standard corporate product. An AI agent is a neural network that does a task step by step: it finds data, uses programs and checks the result. For the market, the question has moved from “do we need agents” to “on which process, and with which number”.
A typical agent for an office role does not know your prices, your CRM, your scripts or when to hand a client to a person. To get a result you need to:
- Connect the agent to your systems.
- Describe the rules.
- Give it a separate account with rights only for its process.
- Agree on a metric before launch.
What should you check now?
Choose one process with a large repeated flow, such as incoming requests, leads or booking. Count how many there are per month and how much time each takes. Find out where the personal data of clients will be stored and on whose servers the model runs. Only then compare platforms and contractors.
A platform gives a model and a constructor, but a process gives the result. A first process can be answers to standard client questions: prices, terms, order status and conditions. The agent answers from your knowledge base and CRM data, and hands hard questions to an employee with the dialog history. Measure the share of requests closed without an operator.