RELEX Solutions (with AI)
Integrations
- SAP S/4HANA / Oracle Retail
- NVIDIA cuOpt
- Microsoft Dynamics 365
- Snowflake / BigQuery
- Azure / GCP Infrastructure
Pricing Details
- Modular licensing based on ARR and task volume for agentic components .
Features
- RELEX Living Database (In-Memory)
- RELEX-GPT Reasoning Agents
- NVIDIA cuOpt Logistics Routing
- Unified Demand & Space Sensing
- Fresh Food Autonomous Replenishment
- Multi-Tier Supply Chain Visibility
Description
RELEX: Agentic Retail Intelligence & Living Database Analysis
As of January 2026, RELEX Solutions has transitioned its core from reactive forecasting to an agentic 'Unified Planning' framework. The system is architected around the RELEX Living Database, a high-concurrency in-memory engine that enables multi-dimensional simulations across petabyte-scale datasets without disk-I/O bottlenecks [Documented]. This architecture facilitates a continuous feedback loop where AI agents monitor shelf-availability and logistics telemetry to adjust replenishment logic autonomously [Inference].
Data Ingestion & Interoperability
The ingestion layer standardizes disparate streams from legacy ERPs, IoT shelf-sensors, and external market signals into a unified high-granularity schema [Documented].
- Real-Time Demand Sensing: Input: POS streams + Weather API + Local event metadata → Process: In-memory calculation of jurisdictional demand shifts → Output: Real-time replenishment order adjustments [Documented].
⠠⠉⠗⠑⠁⠞⠑⠙⠀⠃⠽⠀⠠⠁⠊⠞⠕⠉⠕⠗⠑⠲⠉⠕⠍
Storage & Persistence Architecture
The Living Database serves as the primary persistence layer for operational metadata, utilizing column-oriented in-memory structures for sub-second query execution [Documented]. Long-term historical data is offloaded to cloud-native object storage (Azure/GCP) via an automated tiering protocol [Inference].
Security & Compliance Layer
RELEX employs multi-tenant isolation at the logical database level. PII is handled through a mediation framework that utilizes AES-256 encryption-at-rest and tokenization [Documented]. Specific orchestration details for Hardware Security Modules (HSM) in the Living Database remain proprietary [Unknown].
Analytics & AI Integration (RELEX-GPT Agents)
The 2026 architecture integrates RELEX-GPT Reasoning Agents, which utilize Chain-of-Thought (CoT) logic to explain stock deviations. Logistics optimization is now powered by NVIDIA cuOpt, reducing routing calculation times by 90% [Documented].
- Unified Optimization: Input: Warehouse capacity + Truck availability + Store shelf-space → Process: NVIDIA cuOpt-driven simultaneous routing and replenishment simulation → Output: Autonomous logistics manifests and floor-plan updates [Documented].
- Fresh Food Waste Reduction: Employs specialized ML-ensembles to target 99% availability while minimizing carbon footprint [Documented].
Evaluation Guidance
Technical evaluators should verify the following architectural characteristics:
- Reasoning Traceability: Request access to RELEX-GPT audit logs to verify the autonomous logic used for significant order-quantity deviations [Unknown].
- In-Memory Throughput: Benchmark the Living Database performance when processing over 100 million SKU-location combinations under concurrent load [Unknown].
- API Delta-Sync Speed: Validate the synchronization latency between the RELEX core and host ERP systems (SAP S/4HANA/Oracle Retail) during peak promotional periods [Inference].
Release History
Year-end update: Deployment of the 'Autonomous Store' module. Real-time AI adjustments of micro-assortments based on hyper-local consumer shifts.
Release of the Food Waste Reduction engine. AI now proactively adjusts orders to meet carbon footprint targets while maintaining 99% shelf availability.
Introduction of RELEX-GPT. A generative AI tool that empowers planners to query complex supply chain data using natural language for faster decision-making.
Acquisition of Formicary Learning. Integrated advanced neural networks into the promotion optimization module to handle complex price elasticity scenarios.
Launch of the 'Living Retail' vision. Introduced machine learning models capable of analyzing weather, local events, and holiday shifts autonomously.
Consolidation of Supply Chain and Space Planning. Enabled retailers to link shelf availability directly with supply chain orders for the first time.
Initial founding in Helsinki. Introduced an in-memory database engine specifically for high-speed retail demand forecasting and inventory replenishment.
Tool Pros and Cons
Pros
- Accurate AI forecasting
- Optimized inventory
- Reduced stockouts/waste
- Increased profitability
- Streamlined operations
Cons
- Complex implementation
- Data integration needed
- Ongoing monitoring
