Blog & news 4 min read

Anton Marrero and Serhii Balashuk demonstrated AI solutions on the UnityBase platform

Intecracy Group hosted the offline event Intecracy Solution Day, where Anton Marrero and Serhii Balashuk demonstrated how to integrate AI into the UnityBase platform with a focus on security and scaling.

On October 15, 2025, the Intecracy Group consortium hosted a practical offline event — Intecracy Solution Day. The central theme of this professional gathering was building modern AI solutions on the UnityBase enterprise platform. Leading architects, developers, and IT directors came together to discuss tools for integrating artificial intelligence into corporate systems, with a particular focus on access control, business logic management, and ensuring scalability under heavy workloads.

The event was led by IQusion, an Intecracy Group member.

Modern businesses strive to leverage the capabilities of Large Language Models (LLMs) and generative artificial intelligence to automate routine processes. However, implementing AI at the enterprise level faces serious challenges: from the risk of sensitive data leaks to the complexity of integration with existing registries. The UnityBase platform, designed for building high-performance corporate solutions, offers a systematic approach to resolving these issues, acting as a reliable mediator between artificial intelligence models and corporate databases.

Access Control and Data Security in AI Operations

One of the key topics of the event was data protection. When connecting artificial intelligence to corporate databases, there is a risk that the model might access information not intended for a specific user. Unlike standard solutions where access rights are often configured at the interface level, UnityBase ensures security at the system core using Row-Level Security (RLS) and Attribute-Based Access Control (ABAC) mechanisms.

“When we integrate artificial intelligence into an enterprise environment, we cannot simply grant the model direct access to the database. The architect's main task is to build an interaction where the AI requests data exclusively through the secure UnityBase API layer. The platform automatically applies access control rules for each specific user who initiated the request. This means the neural network will never see or use data for its response that the user themselves does not have legitimate access to,” noted Anton Marrero during his speech.

This approach avoids complex duplication of security logic on the AI agent side. The UnityBase authorization system integrates with corporate directory services (Active Directory, LDAP) and supports modern authentication protocols. Any request from an AI service is accompanied by a user token, ensuring full transparency and the ability to audit system actions.

Business Logic and API: The UnityBase Architectural Bridge

The next block of presentations was dedicated to the practical aspects of business logic implementation. UnityBase allows for the rapid creation of APIs to interact with external services, thanks to its built-in JavaScript engine, which executes code directly within the application server context. This ensures high-speed request processing and minimal latency, which is critical when working with AI agents that require rapid context retrieval.

“UnityBase acts as an intelligent orchestrator. We don't just pass raw data to the model. Through flexible API configuration tools, we form a structured context, clean it, and enrich it with the necessary metadata. This significantly improves the accuracy of AI system responses and reduces computational resource costs. All business logic is described at the platform level, making the system highly adaptable to changes in the architecture of the AI models themselves,” shared Serhii Balashuk.

The platform automatically generates API documentation, simplifying the work of developers engaged in integrating AI services. Thanks to the use of a unified data model, any changes in the structure of UnityBase entities are instantly reflected in the API, eliminating inconsistency between different components of the enterprise system.

Scaling and Performance Under Load

Working with large language models requires significant computational resources and creates an additional load on the infrastructure. The participants of Intecracy Solution Day analyzed the horizontal scaling mechanisms of UnityBase in detail. Thanks to its Shared-Nothing architecture, UnityBase application servers can easily scale by adding new nodes, allowing for the efficient distribution of requests from AI services.

The built-in metadata caching mechanism and optimized database operations through its proprietary ORM allow UnityBase to handle thousands of requests per second on minimal hardware resources. This makes the platform a highly cost-effective choice for deploying corporate AI solutions, where infrastructure costs play a critical role.

In closing the event, the speakers emphasized that the successful integration of artificial intelligence into the corporate sector depends not so much on the choice of a specific AI model, but on the readiness and flexibility of the underlying data management platform. UnityBase demonstrates that enterprise-grade technologies can provide a reliable foundation for building secure, fast, and scalable intelligent systems of the future.