InOrbit unveils ‘RobOps Copilot’ for robot fleets

inorbit.ai reported on April 29, 2024, that InOrbit had unveiled ‘RobOps Copilot’ for robot fleets. We have not been able to verify the primary source, so please refer to the original for details.

InOrbit’s announcement points to the possibility that tasks required to operate robot fleets could shift from analysis requiring specialized knowledge to natural-language conversations. The company says users can query its proprietary data with an LLM and use Slack integration and image analysis through Time Capsule, potentially changing how operations staff work.

What is new in InOrbit’s announcement is its focus on ‘RobOps Copilot,’ which lets users work with robot fleet data in natural language. At the same time, the announcement should be viewed as an extension of the existing RobOps approach: rather than simply visualizing data, it aims to assess autonomy incidents, mission performance and robot health through the same interface and support optimization decisions. The focus here is not on the robots themselves, but on the layer for interpreting operational data and making decisions. Compared with our related article, The Robot Report reports on InOrbit.AI’s OpenRobOps ISO 21423 reference implementation, InOrbit appears to be shifting its focus from presenting a reference implementation for standardization to conversational AI for real-world operations. The former concerned external connectivity and common specifications; this announcement addresses what to interpret and how to make decisions in field operations. Taken together, the two suggest that InOrbit is seeking to cover both robot connectivity standards and the operational UI. An important implication for the industry is that robot operations software is moving from ‘monitoring screens’ toward a layer for questions, answers and interpretation. With Slack integration, operational information could become part of conversations across multiple departments, not just the preserve of dedicated engineers. However, how far decision-making can actually be automated is a separate question; for now, the main functions appear to be supporting explanations and prioritizing issues. Points to watch include which types of robots are supported, how far Time Capsule’s image analysis can go in identifying causes, which features are paid offerings in Standard and Enterprise, and how much the system can reduce decision-making time in the field.

For Japanese robot operations software companies and system integrators, the layer that handles operational data in natural language is becoming a competitive arena, alongside robot control. InOrbit has shown a system for assessing autonomy incidents and mission performance across operations. The question for Japanese companies is how to incorporate their strengths in field implementation and maintenance expertise into conversational operations platforms.