Adagy Robotics recently launched!
"24/7 remote intervention service for your robots"
24/7 remote intervention service for your robots. When your robot fails, it can call out to Adagy over theirAPI. The trained operators will take over and drive your robot back to a stable state, and then resume autonomous operation.
Founded by Rosalind Shinkle and Kathleen Brandes
The Problem 😳
Adagy loves robots, but recognizes that they fail often. For example, a robot might get stuck in tall grass because it interprets the grass in the same way it might interpret a chain link fence – catastrophically impenetrable!
This problem is painful for end users – a farmer who has just spent $100k on an autonomous tractor doesn’t want to stop work to tell the robot 🗣️ “that’s not a fence” 🗣️ every 4 hours.
This problem is painful for robotics companies – they want to be able to sell robots that work, but it can take many years of R&D to get to 99.999% reliability. Robotics companies are used to getting angry calls from customers – or worse, selling a robot that collects dust because it’s too unreliable.
The Solution 👀
Adagy Robotics is a 24/7 remote intervention service for your robots. When your robot fails, it can call out to Adagy over their API. Their trained operators will take over and drive your robot back to a stable state, and then resume autonomous operation. With Adagy, customers won’t even notice that a robot has failed, but robotics companies will be able to learn from the data collected during those failures.
With a safeguard like Adagy in place, robotics companies can deploy and scale robots earlier.
Safety and trust are at the core of their business:
1. They notify site staff before taking control of the robot.
2. They provide incident reports after each rescue.
3. They maintain an auditable log of all failures and the actions taken by their operators.
4. They proactively highlight repeat failures.
Initially, they'll be doing all interventions manually. In the future, they will layer in their own ML models to help the human drivers be more efficient and accurate, to do predictive failure prevention, and offer a software-only ‘supervisor’ to customers.