TB-5559 · REV E · Technical newsheet
Industrial IoT & MonitoringDevice profile
Sensolus puts AI agents on 1.5 million tracked industrial assets
Sensolus adds task-specific AI agents and an MCP connector to its tracking platform covering 1.5 million assets; one agent traced 8 damaged shipments to one carrier.
By Amara Osei4 min read756 words
Features
- Sensolus tracks about 1.5 million assets for 500+ customers via GPS and Bluetooth Low Energy tags.
- Handling Detective agent reviewed six weeks of impact data, flagging 21 threshold-exceeding impacts and linking 8 incidents to one carrier at a border crossing.
- Agents require user approval before acting, and customer data is not used to train AI models.
- The MCP Connector exposes tracking data to ChatGPT, Copilot, Claude and Mistral.
- Customers include Airbus, Pratt & Whitney, DHL, Atlas Copco, RaceTrac and Veolia.

Sensolus now runs AI agents across roughly 1.5 million tracked assets, and one of them recently turned six weeks of impact data into 21 flagged events, eight of them pinned to a single carrier. That is the launch in one sentence: the Ghent-based asset-tracking vendor has added task-specific agents plus a connector that lets external AI assistants query its platform.
The company, which tracks assets for more than 500 customers using GPS trackers and Bluetooth Low Energy tags (short-range radio beacons reporting identity and status to gateways), says the new layer runs on its existing hardware and software. No forklift upgrade. Names on the customer list include Airbus, Pratt & Whitney, DHL, Atlas Copco, RaceTrac and Veolia — logistics-heavy industrials where idle equipment and transit damage are measurable money.
What do the agents actually do?
Sensolus splits the agents by their job, not by their cleverness. Field agents watch the tracking fleet itself: tracker battery levels, scanning setups, and replacement planning for devices approaching end of life. That is the unglamorous work of keeping the sensing layer alive, and it is the precondition for everything downstream.
Operations agents work the asset side. They analyze handling incidents, identify distribution bottlenecks and support fleet optimization — in plain terms, telling you where your returnable containers are sitting still and who dropped them.
The company states two governance facts worth noting. Agents require user approval before taking action, and customer data is not used to train AI models. The second claim matters more than the first; industrial customers treat location and handling data as competitively sensitive, and "your data trains our model" has killed many a pilot.
Takeaway: the agents are assistants with a human sign-off, not autonomous operators.
What did the Handling Detective find?
The vendor's worked example carries more weight than the feature list. Sensolus describes a customer investigation in which its Handling Detective agent reviewed six weeks of impact data from damaged shipments. The agent:
- identified 21 impacts above the customer's defined threshold,
- mapped a concentration of incidents at a border crossing,
- linked eight incidents to one carrier,
- then drafted an email and supporting report for a manager's review and approval.
Note the chain: measurement, geospatial clustering, attribution, and a drafted communication a human approves. The sensor data — accelerometer-triggered impact events on the tags themselves — is the evidence base. The agent correlates; it does not sense. Whether the 21-impact figure would survive an independent audit is untestable from a vendor example, but the workflow structure is credible because the data pipeline predates the AI.
Takeaway: the value here is triage speed on data that already existed and went unexamined.
What is the MCP Connector for?
The second half of the launch is the Sensolus MCP Connector (MCP, or Model Context Protocol, is a standard interface that lets AI assistants pull live data from external systems). Through it, customers can query tracking data from ChatGPT, Copilot, Claude, Mistral and similar tools.
Combined with Sensolus's existing connections to other business systems, an employee can, for instance, examine asset activity next to open orders in an ERP (enterprise resource planning) system. That is the practical pitch: the AI assistant your staff already uses becomes a front end for "where is it and why is it late."
Takeaway: the connector moves Sensolus data into the tools people already have open, rather than demanding a new screen.
The data-first argument
CEO and co-founder Kristoff Van Rattinghe framed the launch in terms any metrologist would endorse. "Our expertise is asset tracking and making the physical world visible," he said. "We want to use AI where it moves our solution and our customers forward. That starts with reliable data. AI can only deliver value if it knows what's happening with assets in the field."
It is the right sequencing. Inference on bad telemetry is still bad inference, just faster. Sensolus's position — a decade-plus of GPS and BLE infrastructure feeding the models — is the moat; the agents are the interface.
Sensolus keeps its U.S. headquarters in Atlanta and its global headquarters in Ghent, Belgium, serving industrial manufacturing, reusable packaging, logistics and environmental services customers.
The open question for buyers: does a human-approved agent layer and a no-training data policy survive contact with your procurement and compliance teams, and can the Handling Detective's border-crossing finding be reproduced from your own raw logs? If yes, this is a workflow change. If no, it is a demo.
via citybiz.co (Source)
Filed under
- ai-agents
- asset-tracking
- fleet-monitoring
- gps-tracking
- bluetooth-low-energy
More from Amara Osei
Show full bio
Senior reporter covering people, launches and day-to-day practice in Instruments, test and measurement at Testbench Report.
124 articles
Application notes
- IPercept raises $16.5M for component-level CNC diagnostics
- BrightAI Raises $97M for Global Infrastructure Inspection AI
- IDTechEx distils three sensor directions from CES 2026
- VEGA Acquires akenza, Bringing IoT Middleware In-House for Process Sensors
- DeGould's AutoCompact DriveThrough Extends AI Vehicle Inspection Beyond the Factory