Real-time analysis of large-scale text data

1. The problem: information overload in intelligence operations

• Intelligence services process vast volumes of data every day.
• Manual analysis is too slow for real-time decision-making.
• Detecting relationships, sentiment and hidden patterns is difficult.
• Military intelligence demands automated, scalable solutions.

2. The solution: ArgusCortex AI for Text

• A natural language processing (NLP) engine processes an unlimited number of documents in real time.
• It extracts more than 15 classes of entity (people, organisations, companies, locations, dates and more).
• It analyses the relationships between entities for deeper understanding.
• It resolves contradictions across documents and enriches them.
• It provides topic modelling, sentiment analysis and key-phrase extraction.

3. Key capabilities

• Entity recognition — people, organisations, aircraft, events, locations, dates and more.
• Geospatial intelligence — maps the location of entities for situational awareness.
• Relationship mapping — links people, organisations, time and events.
• Sentiment and topic analysis — captures the tone and key themes of a document.
• Real-time JSON output — instant access for downstream intelligence systems.
• Multilingual support — Ukrainian, Russian, English, Chinese, Arabic and many other languages.

4. Why now — war as an intelligence challenge

• Real-time intelligence is decisive for making immediate strategic and tactical decisions.
• False data and disinformation spread everywhere.
• Ukraine's intelligence community needs an infinitely scalable, AI-powered tool.
• ArgusCortex AI for Text delivers actionable intelligence instantly — far faster than human analysts and without their errors and omissions.

5. How it works

6. Competitive advantages

7. Real-world use in intelligence operations

• Monitoring enemy communications — identifying key figures, locations and relationships.
• Countering disinformation — detecting false narratives and tracing their sources.
• Investigating financing networks — analysing financial transactions and corporate structures.
• Gauging public sentiment — measuring reactions to political and military events.
• Identifying collaborators and saboteurs — analysing communication and interaction patterns to surface suspicious behaviour.

8. Common program use cases

In this context, a “program” refers to a specific intelligence or analytical initiative run by the client. Examples of the most common programs:

• Monitoring enemy communications — identifying key individuals, places, events and connections.
• Counterintelligence operations — detecting bad actors and saboteurs within internal organisations.
• Disinformation tracking — detecting and tracing false narratives.
• Financial network analysis — tracing financial transactions.
• Maritime and air-fleet intelligence — detecting, tracking and mapping connections and movements.
• Threat assessment — monitoring internal and external political risks and sentiment.

Each of these distinct operational streams is treated as a separate program under our pricing model.

9. Data sources

• Publicly available sources — we have API access to countless open sources.
• Client-provided data — clients can supply their own data for tailored analysis.
• ArgusCortex proprietary data streams — aggregated data from Telegram, WhatsApp and Viber, including reactions, comments and replies.

10. Processing client data

Client data is ingested through a REST API, enabling smooth integration with existing workflows.

11. Flexible deployment options

• Fully hosted (recommended) — no dedicated hardware required. Deployed on AWS or a private cloud platform. Clients can also have a dedicated instance.
• On-premises — roughly 6 GPUs and 3 CPUs are needed for optimal performance.

12. Pricing model

• A fixed annual subscription price per program.
• Our clients typically run between two and ten programs.
• Volume discounts scale with the number of programs.
• Flexible, custom payment options are available.

13. Setup time

A full hosted deployment is completed within 48 hours of signing the contract.

14. Product demonstration

• Sandbox demo environment — we provide hosted trial access to a controlled environment where you can upload a limited number of sample documents and see results in real time.
• Pre-processed sample data — a quick and easy way to get a feel for the product, based on a set of pre-processed sample data and reports.

Conclusion

ArgusCortex AI for Text is an advanced natural language processing and text-enrichment system built on our proprietary artificial intelligence and machine learning models. We are confident that ArgusCortex’s AI/ML models are the best available, and we work every day to refine them further.