Logistics Reply, which is the Reply group company specialized in creative solutions for supply chain execution and warehouse management, on October 5, 2026, announced the LEA AI Agent Authority Model, which is a new practical framework delivered with LEA Reply Dynamic Intelligence to enable organizations to implement AI agents with a suitable level of authority for each task – not maximum autonomy. The fact is that as AI evolves from standard support to live warehouse execution, organizations need a regulated way to determine how much authority agents ought to have in each operational scenario. If there are no clear standards, adoption can stagnate or add more risk to operational groups. The aim is to empower agents with the appropriate power for the task at hand: just because an agent is capable of acting doesn’t mean it should be allowed to.
The LEA AI Agent Authority Model takes on this challenge by combining two dimensions – organizational AI maturity and contextual agent authority. It adds four stages of maturity right from early adoption to mature, to five authority levels – Inform, Recommend, Act, Coordinate and Governed Autonomy. The framework helps organizations understand where standard artificial intelligence can create value today and decide what agents should do, where and with what guardrails, depending on the use case, context and risk.
As operational evidence and confidence are built, authority can then be built.
To make the framework more usable, LEA Dynamic Intelligence provides pre-built agents and agent builder to develop and deploy customer-specific agents. This connects the model’s recommendations on maturity and authority to practical use cases in warehouse operations.
Five pre-built agents automate repetitive warehouse tasks with no custom development are available October 5, 2026. Each has its own data contract and approach to integration, and operates with a degree of delegated authority and oversight by humans suited to its mission and operational context –
- Out of Stock Agent finds the causes of out-of-stock products, classifying them into real shortages and temporary issues, reducing delays and manual checks.
- Labour Distribution Agent detects bottlenecks and estimates effort in order to meet cutoff times and suggests reallocation of workforce when it comes to real-time workload balancing.
- ABC Rebalancer Agent re-calculates ABC classification which based on movement data and generates the reclassification report. If approved, it writes the updated classes to the warehouse management system – WMS item master.
- Docket Scheduling Agent allows planners and carriers to look out for and also book dock-door slots via natural-language conversation by using a guided dialogue that takes scheduling rules into consideration.
- Lost & Found Agent comes into picture when a task is overdue and can use camera input so as to evaluate the environment for causes of delay and identify problems like an item blocking the path of an autonomous mobile robot – AMR.
According to the Executive Partner at Reply, Enrico Nebuloni, “AI is creating enormous expectation, but also genuine uncertainty. Many teams know they want AI but are not sure where it should sit in daily operations or how to adopt it safely. AI maturity is organizational; agent authority is contextual. Our role is to help customers understand where AI can create value today, what level of authority is appropriate for each operational decision, and how to increase that authority safely as trust and evidence develop.”
