Enterprise logistics software company Shipsy has launched Shipsy Brain in beta, a logistics-focused intelligence layer designed to help AI agents make and execute operational decisions. The Gurugram-based company announced the launch on September 1.
Shipsy Brain operates within the company’s AgentFleet platform and supports AI agents across several logistics functions, including document validation, address intelligence, anomaly detection, routing, and settlement management.
“The logistics systems are just systems of record. So, the human is taking the decision and recording it into the system. The intelligence is sitting outside somewhere,” Shipsy cofounder and chief executive Soham Chokshi told in an interview.
Previously, Shipsy’s existing agents relied on frontier AI models. However, Shipsy Brain now uses multiple fine-tuned open-source models tailored for different logistics applications. Chokshi did not disclose the names of the underlying models.
“AI in logistics must understand how shipments, drivers, documents, carriers, contracts, and many other variables interact with each other and then take the right action. Shipsy Brain brings this operational depth to global supply chains. It is built to help enterprises move beyond dashboards and copilots toward AI systems that can reason, recommend, and act within clearly defined business controls,” said Chokshi.
Chokshi explained that general-purpose AI models often struggle to understand industry-specific logistics terminology and documentation. “BOL could mean anything. But we know that it means bill of lading. POD could mean anything, but we know it’s proof of delivery,” Chokshi said.
According to Chokshi, Shipsy designed the specialised models to improve speed and accuracy while reducing operational costs. “Largely, the three points are speed, accuracy, and cost,” he added.
In one live deployment, a major quick-commerce retailer used a Shipsy-powered AI agent to manage orders containing incomplete or suspicious customer information. Previously, drivers could wait for nearly 45 minutes while employees contacted customers to verify whether an order was genuine.
Now, the AI agent contacts the customer, verifies the intent, updates the order, and informs the driver whether to proceed. As a result, Chokshi claimed that the system reduced resolution time to four minutes and helped recover around 30% of the revenue that companies previously lost due to cancelled delayed orders.
However, Chokshi noted that most enterprises still prefer a secure human-in-the-loop framework rather than allowing AI agents to make every decision independently. “Most enterprises today, they want a very secure human-in-the-loop framework,” he said.
Shipsy and its customers jointly establish confidence thresholds for AI-driven actions. Moreover, the company places financial decisions and critical actions, such as order cancellations, behind strict guardrails.
Shipsy said Shipsy Brain draws insights from data associated with more than five billion shipments. The dataset includes over 50 billion operational events, 1.5 billion automated and human decisions, 100 billion GPS pings, and more than 5,000 logistics workflows.
In its internal benchmark, Shipsy said Shipsy Brain achieved an overall field extraction score of 82.2%. The company compared this with 63.4% for Gemini 3.5, 62% for Gemini 3, and 62.4% for Gemini Pro.
Shipsy Brain also scored 92.4% in logistics-domain knowledge, compared with 45.9%, 38.6%, and 45.9%, respectively, for the three Gemini models, according to the company.
Chokshi said the company tested the models using real documents from field operations. The team provided each model with the same context and evaluated their responses against verified answers using an identical scoring script. However, he did not disclose the size of the test dataset.
Shipsy currently serves more than 150 enterprise customers. Looking ahead, Chokshi said the company aims to deploy at least four or five AI agents across half of its customer base by the end of the year.
