Bluesonde deployed two buoys at Newtown Creek Wastewater Treatment Facility, generating 1,500 real-time data points across five water quality parameters and detecting seven combined sewer overflow (CSO) events, including how long the creek takes to return to safe conditions after an overflow. DEP identified the strongest near-term use case in continuously tracking the dissolved oxygen levels, which would require Bluesonde to develop a subsurface sensor. Continuous dissolved oxygen readings could replace twice-weekly manual sampling and help pump in only as much air as needed, reducing energy costs associated with aeration. If a subsurface sensor that could continuously monitor dissolved oxygen was deployed across five Newtown Creek sites, it’d be roughly 25% cheaper than the current manual sampling costs, saving over $300,000 within the first year.

Cyvl scanned an area in Astoria, Queens, previously reviewed by internal teams, and analyzed more than 700 DEP assets across three use cases. Use case #1 - work order validation: Cyvl automatically detected and validated 311-generated maintenance requests using street-level scan data, achieving approximately 25% automated detection rates and projected savings of $500,000 per year by eliminating unnecessary field visits. Use case #2 - asset recovery: Cyvl identified ~1.7% manholes in the scanned area had been partially or fully paved over during road resurfacing. DEP estimates that recovering buried manholes across the city during each paving cycle would cost between $8 million to $32 million. Sequential Cyvl scans could enable DEP to flag and recover these assets before they beco emergencies. Use case #3 - street flooding investigations: Cyvl showed that physical measurements previously requiring hours of field work could be pulled virtually from scan data, enabling an estimated 7 to 10 times increase in yearly investigation capacity and projected savings of around $1 million in field costs.

Nyad deployed their platform with Newtown Creek and 26th Ward wastewater treatment operators. This gave process engineers an AI chat tool for real-time decision support and a computer vision model to analyze microscope images of activated sludge, receive AI-generated diagnostics, and discuss targeted operational recommendations. Operators logged 473 messages across 55 conversations and uploaded 198 microscopy images, with average response times under 60 seconds. At scale, Nyad could give engineers near-instant insight into biological conditions that typically take days to verify, while preserving plant-level institutional knowledge.

Sunfish deployed its autonomous underwater vehicle at New Croton Dam for over 7.5 hours across two days, inspecting a wall section and intake structure to detect zebra mussels, freshwater mollusks that can compromise water quality and operations.  Unlike traditional methods, Sunfish can collect data with no crane setup, drain-down, or workers entering the water. The high-resolution data from the inspection helped to detect zebra mussels and validated measurement accuracy to within 8%. The more promising use case is around using Sunfish to conduct inspections of deep tunnels and pressure aqueducts inaccessible to traditional remotely operated vehicles. Sunfish is currently developing a product that would be able to operate under larger flow rates and pressure heads to make this possible. Once developed, Sunfish projects this approach to be roughly 60% cheaper than conventional methods.

Teamsolve ran two projects with different DEP groups. On the wastewater treatment side, the platform aggregated five years of chemical cost and consumption data across six facilities to forecast future chemical costs under different scenarios to support budget planning. On the water supply side, it analyzed maintenance work order data to identify assets with recurring failures and overdue maintenance. At scale, Teamsolve's value sits in the ability to pull together disconnected data and turn this information into a single, usable picture, whether that’s chemical costs, equipment health, or other operational data.

TrelliSense deployed a sensor at Wards Island Water Resource Recovery Facility that tracked methane, a potent greenhouse gas that forms as the plant processes solid waste and can escape from the tanks. The sensor monitored emissions from digesters and thickeners in real time using laser beams that continuously scan the air across the facility rather than relying on periodic surveys by staff with handheld detectors. The TrelliSense sensors identified variation in methane leak rates, and this data helped to tie elevated emissions to specific operational events. The POC established continuous monitoring as a strong candidate for compliance with upcoming NYS DEC Regulation 253, which will require wastewater facilities to report annual greenhouse gas emissions, including methane, starting in 2029.