Quadrupling the speed of sewer inspection with artificial intelligence.

Quadruple the speed of sewer inspection with AutoCode



Every year 860 billion gallons of raw sewage overflows into our streets and waterways. This problem is getting worse because many pipes were installed after the 1972 clean water act but are now at the end of their useful life. The EPA currently fines US cities about 1 billion dollars a year for overflows.

To mitigate this risk, average cities in the US inspect about 240K feet a year at about $2/foot, which costs them $480K annually. The process is very slow as specialty trucks (a $350K investment) with 2 operators lower tethered cameras into manholes and slowly drive the cameras, watching the video footage so they can label pipe defects manually using a labeling system called Pipeline Assessment Certification Program—or “PACP”—coding. For example, the presence of tree roots = “RFJ.”

This slow process means that cities can cover a small percentage of their cities each year, leading to few defects caught and more overflows. During inspections the camera is sitting still 2/3 of the time indicating lots of wasted inspection time.


Meet AutoCode.

AutoCode is software that cities can install on their trucks to speed up the inspections. The software automatically recognizes the defects in the video stream and fills out the “PACP” database so the operators no longer have to stare at the screen and manually fill out the database. This gives them the opportunity to do more meaningful work.

AutoCode speeds up inspections 4 times their current speed which lower’s costs from $2/foot to $0.75/foot.

AutoCode will give cities the ability to inspect much faster and more frequently.

AutoCode will enable cities to move from expensive reactive maintenance plans to lower cost proactive plans because they will be able to find and fix small defects before they become expensive catastrophes. This will save every city millions in repair and maintenance.

mineral deposit.png