Skip to content 
Moscow, January 26.
According to the information TASS citing the press service of the Baikov Institute of Metallurgy and Materials Science of the Russian Academy of Sciences, the scientific team of IMET RAS in partnership with the company Ensec AI has developed the SyncPoint platform. The report emphasizes: it is neither a separate chatbot nor a «document search», but a platform layer that integrates disparate corporate data and artificial intelligence tools into a single loop to support managerial and scientific and technological decisions.
- - Why it's more than just another AI assistant
- - What is claimed to be the core of the platform
- - What the SyncPoint product's public documents confirm
- - Industrial circuit and partnership with Rostelecom«
- - Status of domestic software
- - Significance for expertise and subsoil use
- - Materials for independent study
Why it's more than just another AI assistant
Chatbots, corporate search engines, assistants for lawyers and HR departments are already commonplace in large companies. But SyncPoint developers emphasize the main deficit: these are almost always point solutions. What was missing was a platform layer that can collect data from multiple systems and turn it into a single intelligent space where search, analysis and automation work in concert.
Return to contentWhat is claimed to be the core of the SyncPoint platform
- - A single point of interaction with corporate data. The platform integrates documents, knowledge bases and pluggable information sources.
- - Hybrid search and context-aware answers. Full-text search, semantic search and generative AI work together.
- - Intelligent agents. The agent layer is designed not only for answers, but also for performing tasks: analyzing, comparing versions, preparing summaries and reports.
- - Roles, access rights and working in the loop. Claimed compliance with access restrictions and the ability to deploy to an organization's infrastructure.
What the product's public documents confirm
Interface and basic script
The user manual describes the interface structure: documents and knowledge bases, chats, agent store, search, favorites and personal cabinet. The basic scenario is built around uploading documents, starting parsing, waiting for readiness status and further work in chat with connecting an agent for a task. The workspaces include the roles of owner, editor, and guest.
Searching and indexing
A document processing pipeline with structure-aware partitioning into semantic fragments is described. Search is declared to be hybrid: full-text plus semantic on embeddings plus metadata. The index is updated when new documents are loaded and can be re-indexed on a schedule.
Models and connectivity
The platform claims to connect large language models and embedders via an OpenAI-format compatible API. Practical sense: SyncPoint is positioned as a platform orchestrator over models, not as a developer of its own LLM.
Agents as the execution of action chains
The functional requirements list a set of agent tools: web search via Yandex API, parsing web pages, executing Python in an isolated environment, building graphs, working with docx and excel, getting data via Moscow Exchange API. This means that agency is stated as an orchestration of tools, not just prompting.
The point for the customer: the documents allow us to see that we have a web platform with storage, index, parsing, chats and agents, not a «magic bot». This is the right level of transparency for a serious implementation.
Return to contentIndustrial circuit and partnership with Rostelecom«
The project is supported by industrial cooperation. In 2025 «Rostelecom» and DeepTech company ENZEK signed a cooperation agreement at SPIEF-2025. The official announcement states that the parties plan to jointly develop SyncPoint as a tool for integrating AI into the corporate environment.
Return to contentStatus of domestic software
The SyncPoint platform is included in Unified Register of Russian Computer Programs and Databases. The public registry card shows registry entry #31094 with a date of 12/10/2025.
Return to contentSignificance for expertise and subsoil use
Expertise in subsoil use is based on evidence: the version of the report, the source of the figures, the consistency of sections and appendices, and the reproducibility of calculations. SyncPoint-level platforms potentially reduce the most costly steps: searching through arrays of materials, reconciling versions, gathering context and preparing consolidated analytical materials.
Important: this is not an «expert replacement». It is an infrastructure that removes the manual routine around data and accelerates getting to the heart of the solution, while keeping the responsibility for the output with the expert.
Technical dossier for specialists
Search outline: hybrid schema full-text plus embeddings plus metadata, index update on load and scheduled re-indexing.
Agent outline: web search, parsing, Python execution, office file handling and graphing tools.
Metrics: the requirements mention response quality and performance metrics, including faithfulness, context recall, RPS and p95 latency, which are correct to test in a pilot on customer data.
Return to content