Alvys Foundry ships with 20 plus agent templates for detention, paperwork, and rate audits, gated by an SOP upload and a simulation test before live use.
Alvys has put AI agents inside its freight management software so brokers and carriers can configure them to file detention claims, audit freight invoices, and process bills of lading inside the system they already run. The new product, Alvys Foundry, is the agentic layer that sits on top of Alvys's existing transportation management system (TMS), the back-office software trucking companies and the middlemen who arrange loads use to plan routes, dispatch drivers, and settle paperwork.
Foundry ships with more than 20 pre-built agent templates, according to Alvys's announcement on PR Newswire. The templates target six recurring jobs: detention processing (fees a carrier can claim when held at a shipper beyond an agreed free time), document handling for rate confirmations, bills of lading, and proofs of delivery, rate audits against agreed tariffs, shipment tracking and check calls, asset compliance checks on authority, insurance, and safety, and claims management.
Customers can upload an existing standard operating procedure or describe a task in plain language, and Foundry generates a workflow for human approval, with simulation-based testing before the agent goes live. The platform runs against Alvys's existing footprint of more than 120 integrations and native EDI (electronic data interchange) connections to hundreds of shippers, so agents act on the same data the brokerage already touches.
Alvys CEO Nick Darman framed Foundry as a response to customers adopting AI tools that did not fit existing operating procedures, per FreightWaves. The launch extends earlier assistive features, including load-margin insights, weather and cargo-theft alerts, and auto-creation of loads from rate confirmations, first reported by Artificial Intelligence News.
What remains unproven is whether the agent execution layer will absorb real workload. The launch is a vendor self-description, and no independent customer, analyst, or adoption-metric data is available yet.