Samsara connects fleet data with ChatGPT, Claude and other AI tools

Samsara is giving customers a way to connect their fleet and operational data directly to AI tools such as ChatGPT, Claude, and Microsoft Copilot, among others. The software company announced on Sept.

22 that its new Samsara Model Context Protocol (MCP) allows authorized users to query live information from their Samsara accounts through third-party AI applications without building a custom integration. More than 40 read-only tools are available through MCP at launch.

They allow AI applications to retrieve information about vehicles, drivers, safety events, hours of service and other operational data. Developers can also use the tools when building their own AI agents. A FreightWaves report published on Sept.

23 detailed how Pam Polyak, owner of Wisconsin-based Polyak Trucking, used Claude to analyze data from her Samsara system. She asked the AI tool to identify drivers spending more than 30 minutes on duty without driving while at the company’s terminal.

The analysis took about three minutes and identified a driver who was spending more than 110 minutes per shift in the yard without driving. Polyak said the company recovered $53,000 in labor costs after identifying the issue.

She also described using Claude to compare toll costs between drivers running the same lanes, which uncovered a $50-per-load difference. The connection uses the permissions already assigned to each user in Samsara. If a user cannot access a particular piece of information in Samsara, the connected AI application cannot access it either.

Samsara said each request is tied to the individual user making it, rather than relying on shared API keys or service accounts. The tools are read-only, so an AI application can retrieve information from Samsara but cannot change records through MCP. Samsara said customers can also use its data alongside information from other business systems.

The company cited examples involving human resources, finance and procurement, including comparing driver hours-of-service records with time and attendance information, analyzing fuel spending alongside mileage and idle time, and reviewing equipment utilization against rental agreements.

The launch gives companies a way to use fleet data alongside information from other parts of the business. “Our customers are building AI into more of their business, and decisions in areas like finance and HR need to reflect what’s happening in the field. Samsara provides that operational understanding.

With MCP, customers can bring their own Samsara data into the systems and AI tools those teams already use, connecting the work of running their operations with the work of running their business,” O’Driscoll said. Samsara said its broader ecosystem includes more than 400 integrations, APIs and developer tools.

MCP is built on the same live operational data layer as Samsara Assistant. Samsara also introduced Agent Studio at its Beyond 2026 customer conference. While Agent Studio allows customers to create agents within Samsara, MCP connects Samsara data to third-party AI applications and custom agents.

“With MCP connections already in place across our other major business systems, we can now bring Samsara’s operational context into those same workflows. That lets our operators focus on solving business problems instead of maintaining yet another custom integration,” said Ken Apple, CTO at Mariner Logistics.

Samsara MCP is now available to customers at no additional charge. The company said future MCP costs will be based on monthly usage. Why this matters AI tools can already analyze fleet data when companies build their own workflows, but doing so can require custom integrations.

Samsara’s MCP gives customers a standardized way to bring live fleet information into third-party AI applications, potentially making it easier for fleet operators and other business teams to use operational data in existing AI workflows. The post Samsara connects fleet data with ChatGPT, Claude and other AI tools appeared first on FreightWaves.