Supply chain teams can now use natural-language questions to investigate logistics performance through a new integration between project44 and Snowflake. The connection combines project44’s logistics data and business semantics with Snowflake Cortex AI, supporting its Mo AI supply chain analyst.
Mo is designed to do more than generate conversational responses. Its agent framework interprets a question, applies customer-specific definitions, retrieves the relevant operational data and recommends possible next steps. Users can ask about delays, carrier performance, working capital or disruption exposure without waiting for a separate report or analysis.
Context is critical for reliable automation
The integration draws on project44’s logistics ontology and network data, which covers approximately 1 trillion shipment records accumulated over the past decade. It also uses real-time milestones and predictive estimated arrival times. The company says this shared structure helps Mo distinguish how individual customers define concepts such as on-time delivery, transit time and underperformance.
That distinction matters in industrial logistics. One shipper may measure transit from cargo-ready date, while another starts at pickup. Contract terms, transport modes and customer exceptions can also change the meaning of a seemingly simple question. Mo applies those rules before retrieving an answer.
Snowflake Cortex Analyst translates the interpreted request into SQL and obtains results directly from the customer’s project44 data in Snowflake. This grounds the response in structured records rather than figures reconstructed from a language model’s general knowledge. More information on project44’s Mo platform is available from the company.
From analysis to operational action
In internal testing, Mo answered all 10 verified questions correctly, compared with two correct answers from a general-purpose large language model. It returned the same answer in all 10 repeated tests, while the comparison model did so twice, and responded 2.5 times faster. Project44 says 210 customers now use Mo, which has handled more than 50,000 questions.
The reported financial impact includes a $6 million reduction in detention and demurrage exposure for one enterprise customer, $200,000 released from working capital through a single query for another, and nearly $750,000 saved through a disruption-management initiative. For industrial companies, the development points to a practical role for AI: turning complex logistics data into consistent decisions that can be acted on by a wider group of employees.
