EverestLabs Puts AI Agents Near Recycling Controls, but Keeps Operators in Charge

EverestLabs is promising an AI layer that can spot losses on a recycling line and prepare equipment changes. But Navigator’s central design choice is restraint: the system may diagnose and recommend, while plant operators must approve every control adjustment.

The company launched Navigator on August 24 as a system for materials-processing and recycling facilities. It monitors material moving through a plant, diagnoses operating problems and prepares changes to equipment settings for approval. EverestLabs says the product can connect to legacy controls without replacing existing lines or requiring extensive retrofits.

Navigator combines edge computer vision, vision-language models and four specialized software agents. Cameras classify objects on conveyor belts; the vision-language layer is intended to interpret changes in material composition or contamination. The agents cover data queries, vision alerts, equipment settings and fleet health through a conversational interface.

• Data-query agent: handles questions about plant data.
• Vision-alert agent: surfaces issues detected in material flows.
• Equipment-settings agent: prepares proposed changes to settings.
• Fleet-health agent: covers the condition of connected equipment fleets.

That setup is designed to make the interface operational rather than purely analytical. An operator could ask where valuable PET plastic is being lost and receive a diagnosis tied to a line, shift or piece of equipment, followed by a recovery plan or equipment preset. The final step remains with the operator, including when a change is passed through a SCADA industrial-control system.

EverestLabs estimates in its launch release that Navigator can increase facility capacity by 20% to 30%. Its product page uses a broader claim of 20% to 40% higher throughput. Those are company estimates, not independently audited results, and the differing ranges leave the return dependent on a facility’s starting utilization, equipment and incoming material.

The company’s product page also estimates that a 400-ton-per-day facility could generate $2 million to $4 million in additional annual revenue, and it projects roughly $1 million in annual savings from preventing downtime. It says typical facilities lose $100,000 to $800,000 in recoverable material to landfill annually. Whether those figures translate to a particular site will depend on its operating baseline and whether staff implement the recommendations.

Caglia Environmental is an early Navigator customer. EverestLabs also named Schneider Electric and sorting-equipment maker Pellenc ST as integration partners, relationships that could help Navigator reach the controls and machinery it needs to move beyond a plant-data dashboard.

EverestLabs says Navigator builds on systems deployed across more than 200 lines and more than 10 million belt hours. Those deployment figures are self-reported and undated. The more consequential proof point is still ahead: whether its recommendations improve throughput, downtime and material recovery in operating plants while preserving the approval workflow the product promises.