Planet labs now have the ability to map the entire planet every day at a 3m resolution. So object detection applied on these images can be a quite efficient. I still wouldn't call this AI as we are talking mostly about supervised learning but it's a highly practical real world use case. CrowdAI and the like are on this path already.
It is battle tested by me and I can say it is surprisingly accurate.
We would not fit his definition of a vertical specific startup. (We have also been around a while though) The bulk of what we do is time series.
Applications we do for real paying customers include:
Detecting theft of power on the raw grid
Online payments fraud
Detecting people stealing from the telco network
Detecting faults in assembly line machines
Detecting computers about to fail
Detecting root cause of dropped calls
Kind of researchy, but we've also done robotics with RL to teach a robot to learn an obstacle course.
The problem is that though the above may currently be a profitable business, I can't see how you will generate a "moat" and mature into a monopoly or a big market share in a vertical.
Skymind is a weird mix of joint ventures in asia, US and Chinese investment. We're not looking for an exit anytime soon. We are looking to build a big business though.
The "moat" is a land and expand strategy. There is lockin with our tooling. It's a standard on prem play. We're hard to get rid of once you install us. We help in house teams compete with external vendors. Interest actually aligns there. Happy to elaborate a bit otherwise.
Predicting real estate opportunities (which houses are about to sell) predicting energy usage for thousands of sites in real-time, predicting churn, predicting Ad-Tech prices and fraud in real-time, predicting anomalies in seismic radar scans, customer segmentation and recommender engines.
Those are examples of real world paying customers using Deep Learning