At ReFi Financial Solutions Agency, AI Agents Write Reports
ReFi Financial Solutions Agency has deployed AI agents to automate how deal analysts interact with the management system.
The Financial Solutions Agency (FSA) helps companies and individual entrepreneurs secure financing through banks, leasing companies and private investors. Every day, the agency’s professionals support dozens of multistage deals, where incorrect data or a delayed decision can come at a high cost. Artificial intelligence has helped reduce the impact of human error.
A Matter of Minutes
Previously, employees had to prepare a report on every deal at the end of the workday, manually filling out dedicated cards to record its status. “In practice, this could take 10–15 minutes for a single complex deal, and with a large number of projects, it turned into a separate evening task. Now an employee simply dictates a voice message, which takes just a few minutes on their end,” said Renat Khilazhev, CEO of ReFi Financial Solutions Agency.
The voice message is sent in free-form language not to a manager but to an AI agent through a Telegram bot. The agent structures the information, analyzes the outcome, flags any problems, formulates a recommendation for the next required step and automatically updates the CRM system’s records. By morning, the management team already has a completed report on all active deals that require oversight. This has significantly reduced the workload for both analysts – each report now takes less than five minutes – and managers, who no longer have to dig through comments and correspondence to determine whether the data is up to date. The business impact of intelligent automation is also reflected in greater process transparency and faster decision-making.
An Agent With Limited Access
Seeking a balance between system autonomy and controllability, ReFi tested several different approaches and then improved its infrastructure after selecting one of them. The company created a protected environment for AI agents on a dedicated server and blocked the algorithms from accessing confidential information. “Deploying AI is not so much about choosing the smartest model as it is about building the right architecture around it,” Renat Khilazhev said.
The ability to address information-security requirements and provide deep industry-specific adaptation will determine how widely agent-based systems are adopted in the future. Specialized industry agents are emerging as one of the most promising directions. In finance, they can assess credit risks and prepare structured analyses; in manufacturing, they could analyze technical documentation and support engineering decisions; and in the public sector, they could process large volumes of requests and regulatory documents.
From Chatbots to Digital Employees
The path to digital employees in Russian companies has been relatively short. In 2022, large businesses mainly used AI for data analysis, customer interactions through chatbots and document workflow automation. In 2023, generative models and machine-learning systems began to be adopted more actively, helping automate customer service and process large volumes of information. Digital assistants delivered measurable business benefits, driving demand for these solutions. By 2024, about 40% of large companies were already using AI in business processes. In 2025, the market shifted toward agent-based systems capable of carrying out sequential chains of actions, and by 2026, these solutions had become part of management infrastructure.
Intelligent Automation
Digitalization is giving way to intelligent automation, in which some analytical and oversight functions are handed over to algorithms. In the coming years, the number of deployments of agent-based systems is expected to grow in industries that handle large volumes of data and complex interaction chains – finance, logistics, telecommunications, manufacturing and government. This will require not only specialized IT products but also regulatory adaptation, including transparent mechanisms for verifying AI actions, particularly when sensitive data is involved. Experts note that the adoption of agent-based systems will change industry business standards.
