Proper wastewater management is essential for safeguarding public and environmental health. Aging infrastructure, increasingly stringent regulations, resource constraints, and rapid urban development place growing pressure on utilities to operate more efficiently and proactively. To meet these challenges, utilities must understand wastewater network behavior and continuously monitor wastewater composition to detect changes early and intervene effectively. Continuous, high-quality data collection is a prerequisite for understanding system dynamics and supporting timely, informed decisions.
Kando’s Wastewater Intelligence system enables continuous, real-time monitoring of wastewater quality and targeted sampling across the collection network. The hardware component provides reliable, high-frequency data on wastewater behavior, which is then validated and analyzed by Kando’s AI and machine learning algorithms to detect deviations, identify patterns, and assess risk. When high-risk wastewater quality events are detected, the system automatically alerts users via the dashboard, email, or Telegram, ensuring rapid awareness and response.
To support daily operations and improve accessibility to insight, Kando has developed STREAMi, an AI-powered voice and chat assistant embedded within the Wastewater Intelligence system. STREAMi provides natural language access to validated, algorithm-generated insights, trends, and sampling information, enabling pretreatment, compliance, and operations teams to interact with the system efficiently, whether in the field or at the desk. By simplifying access to complex analytics, STREAMi supports faster understanding and more confident decision-making, while expert services ensure effective adoption and use.
Together, the system enables utilities to:
Improve overall wastewater quality through early detection and intervention
Strengthen source control and pretreatment effectiveness
Reduce operational risk and prevent unplanned WWTP disruptions
Support regulatory compliance with consistent, data-driven evidence
This paper provides an overview of the Kando Wastewater Intelligence system delivered as Data as a Service (DaaS), with particular emphasis on the hardware component, including Kando’s proprietary, contactless A-EYE sensors that enable continuous data collection in challenging sewer environments.
Figure 1 Kando DaaS Wastewater Intelligence Features & STREAMi
Kando wastewater intelligence system components
Kando wastewater intelligence solution is a comprehensive DaaS that seamlessly integrates software, hardware, and expert services, all powered by cutting-edge AI and ML technologies. This integrated approach enables Kando to deliver real-time wastewater intelligence through Kando’s User-friendly dashboard.
Figure 2: Kando Wastewater Intelligence DaaS is composed of Software, Services, and Hardware
As depicted in Figure 2, Kando Wastewater Intelligence DaaS is composed of three key components:
Kando’s Wastewater Intelligence system is delivered as a comprehensive Data as a Service (DaaS) solution that integrates hardware, AI-driven software, and expert services. Together, these components enable utilities to transform raw wastewater data into actionable, system-level intelligence, accessible through Kando’s dashboard and supported by expert guidance.
Software (AI and Analytics)
The software component forms the intelligence layer of the system, processing real-time and historical wastewater data using advanced AI and machine learning models. These algorithms validate data, detect wastewater quality events, classify sources, and generate insights that support operational and strategic decision-making.
Hardware (Data Collection)
The hardware component enables continuous, real-time data collection through the deployment of contactless A-EYE sensors, controllers, and automated samplers. These devices provide the high-quality data required for accurate modeling and insight generation, even in harsh sewer environments.
Expert Services (Adoption and Outcomes)
Expert services support the deployment, operation, and adoption of the system. Wastewater experts, data scientists, and field teams ensure the hardware is installed effectively, the algorithms remain accurate, and insights are aligned with utility workflows and KPIs. Expert services do not generate intelligence independently, but ensure that AI-driven insights are trusted, understood, and applied across departments.
The hardware component captures real-time wastewater data, while the software continuously analyzes and contextualizes it. Together with expert services, this integrated approach ensures utilities can detect wastewater quality issues early, understand their impact across the network, and act proactively.
Figure 3 Kando wastewater intelligence components and data flow
The system consists of the following components:
Component | Attribute | Functionality |
Kando’s proprietary contactless A-EYE sensor, pH, ORP, EC | Measures anomaly levels in wastewater flow | |
| Data logger | Records the data collected by the sensors and transmits it wirelessly to the software component |
| Automatic event-triggered sampler | Collects wastewater samples when a wastewater event is detected |
AI/ML powered software | Automatically detects wastewater quality events through its AI & ML models (such as anomaly and pattern detection) | |
| Dashboard | Presents algorithm-generated wastewater intelligence, including validated sampling data, parameter-level graphs, Pollution Index (PI) scoring, hierarchical network views, and system-wide wastewater quality trends, enabling teams to assess conditions, prioritize risks, and act on insight rather than raw data. |
| API | The API enables seamless sharing and utilization of wastewater insights across various components of the utility system, including the SCADA team, BI platforms, and GIS team, thereby enhancing data analysis capabilities |
| Cyber Security | Data security and privacy are built into every layer of the solution, with robust cybersecurity frameworks and compliance measures in place to protect sensitive utility information |
Deployment | Analysis of the network through data models (NetFix for example, GIS-based assessment tools, etc) | |
| Maintenance | Maintain the hardware in the collection systems and optimal data collection |
| Wastewater experts | Accessibility to wastewater experts to train and support the utility team in utilizing the wastewater intelligence in their daily operational strategies |
| Data Scientist | Continuously keeps the software and algorithms up to date with the latest developments |
| STREAMi | Kando’s voice-powered wastewater intelligence assistant empowers management, field and operations staff to interact with the platform through natural language commands. |
Measuring Contamination with Kando’s Hardware
As part of its hardware component, Kando integrates A-EYE sensors, a real-time non-contact fluorescence-based sensor, and a water level sensor. These sensors are designed to provide cleaner data while ensuring scalability and reliable, real-time wastewater monitoring. The A-Eye sensor stays above the wastewater, avoiding contact with contaminants, thereby reducing maintenance needs and enhancing data accuracy.
Figure 4 Kando's equipment to retrieve real-time data installed in a manhole
How A-EYE sensors work
A-EYE sensors utilize fluorescence spectroscopy to detect and quantify pollutants in wastewater.
Fluorescence spectroscopy relies on molecules that absorb light at a specific wavelength, which excites their orbiting electrons to a higher energy state. As the electrons return to their ground state, they emit light at a characteristic wavelength(Figure 5). The A-EYE sensor relies on the natural fluorescence of the pollutants themselves to detect and quantify their presence in wastewater.
The optical channels of the A-Eye sensor (Figure 6&7), in combination with machine learning models and algorithms, enable continuous and rapid detection of pollution events involving organic matter, heavy lubricants, fuels, mineral oils, and biological contaminants, thereby providing real-time insights into wastewater quality. Detergents are detected with lower sensitivity.
Figure 5 The A-EYE sensor and its features and Installment
Figure 6: A-Eye fluorescence channels
Additionally, the A-Eye sensor utilizes supplementary optical effects such as light scattering and absorption to enhance pollution detection capabilities.
The following process describes how an A-EYE sensor obtains data:
The sensor emits a UV light beam of a specific wavelength onto the wastewater flow every 5 minutes.
In response to the light, the organic matter in the wastewater emits low-energy photons in a phenomenon called fluorescence.
The sensor detects the fluorescence and measures its intensity.
The sensor forwards these measurements to the AI/ML-powered software component, which uses them to detect wastewater quality events.
Figure 7 Detecting events based on excitation and emission wavelengths
By combining this sensor data with machine learning algorithms and domain knowledge, the system continuously monitors wastewater quality and detects contamination spikes.
Real-Time Data Processing and Analysis
Following the acquisition of raw signals from A-Eye, machine learning models and algorithms process the data to:
Detect anomalies and classify pollution events.
Provide real-time insights to stakeholders such as utility managers and environmental regulators.
Assist in protecting wastewater treatment plants, improving network efficiency, and promoting sustainable wastewater management.
Water Level Sensor and Signal Correction
The water level sensor is a complementary component of Kando’s system. It is used for:
A-Eye signal correction, ensuring accurate fluorescence measurements under varying wastewater conditions.
Flood event detection, alerting utilities to potential overflow risks within manholes.
The Process: From Data Acquisition to Pollution Event Classification
Kando’s solution follows a structured approach:
Data Acquisition – Collecting real-time sensor data, lab samples, and manhole specifications.
Preprocessing – Cleaning raw data for enhanced accuracy and reliability.
Feature Extraction – Identifying key data attributes that aid in pattern recognition.
Model Training – Using historical data to train AI models to differentiate between normal and anomalous conditions.
Anomaly Detection – Identifying deviations that signal potential pollution events.
Pollution Event Classification – Categorizing anomalies based on predefined pollution types.
Reporting – Delivering event-based wastewater quality insights to stakeholders.
Customized Support – Providing tailored reports for utility managers, regulators, and industries.
By leveraging advanced AI and machine learning, Kando ensures that pollution events are detected and classified in real-time, empowering utilities with actionable intelligence to enhance wastewater quality management.
How A-EYE sensors differ from traditional EC and pH sensors
A-EYE sensors measure: | Traditional sensors measure: |
Fluorescence intensity | EC and pH |
Unlike traditional EC and pH sensors, which provide basic chemical measurements, A-EYE sensors focus on fluorescence intensity to determine pollution sources. This advanced sensing approach allows for:
Rapid and precise identification of organic and industrial contaminants.
Continuous remote monitoring, reducing fieldwork requirements.
Early detection of pollution events, enabling timely interventions.
A-Eye sensors can measure wastewater with a water level height exceeding 4 cm, ensuring high sensitivity across varying flow conditions.
By integrating real-time, non-contact monitoring with ML-driven analytics, Kando provides a scalable, high-precision solution for wastewater pollution detection and prevention.
Building the software engine to detect wastewater quality events
Measurements collected by sensors, combined with historical data and other data sources, are used by Kando to build its AI/ML-powered software component that accurately detects wastewater quality events.
In the past three years, Kando collected additional data from a network of research stations located in different manholes. These stations provided data on fluorescence as measured by A-EYE sensors, as well as on the following parameters:
chemical oxygen demand (COD)
electrical conductivity (EC)
pH levels
Data also included lab sampling results. This data was used to create an additional Machine Learning (ML) model that develops rules for detecting wastewater pollution events.
Research findings: A-EYE sensors lead to the detection of more wastewater quality events
Along with collecting data, Kando also conducted research comparing water-touching sensors and A-EYE remote sensors for event detection. Various hardware components, including A-EYE sensors with multiple channels, EC and pH sensors, an oxidation-reduction potential (ORP) sensor, water level sensors, a data logger, and an autosampler, were used to collect continuous data every 5 minutes. Additionally, lab results were obtained from samples collected following predetermined rules.
The research demonstrated the system’s ability to correctly identify pollution events.
When used with real-time A-EYE sensors’ measurements, the ML model successfully detected pollution events, as confirmed by lab results. On average, using the set of measurements from A-EYE sensors resulted in the detection of 47% more events than were detected when using the set of measurements from traditional sensors. Furthermore, confirmed by lab sample results, ~80% of pollution events detected were positive.
Figure 8. Comparison True Positive pollution event detection EC & pH measurements (blue) versus the A-EYE sensor measurements (pink)
Area type | Improvement in accuracy |
Collectors | 41% |
Factories | 64% |
Average | 47% |
Measurements collected by A-EYE sensors help detect wastewater quality events
A-EYE sensors do not measure exact concentrations of any particular pollutant. However, the measurements of fluorescence intensity are used by AI/ML-powered software component that interprets the measurements to determine the wastewater quality event.
Kando plans to enhance the software's ability to determine the severity and types of events (metals, FOG, and more).
Durability and scalability
In terms of durability and scalability, A-EYE sensors outperform traditional sensors in several ways. They are not submerged in wastewater, ensuring longer lifespans, resistance to wastewater flow impact, and freedom from obstructions by solids or fats, ultimately reducing maintenance requirements. These sensors can be distributed more extensively throughout the network, providing a clearer and more comprehensive picture of wastewater quality.
Continuous monitoring
A-EYE sensors collect data every 5 minutes and record it in data loggers, which, in turn, transmit the data wirelessly to the software component system every 15 minutes.
Figure 9 Data logger installed in a manhole
With traditional sensors, monitoring is often interrupted for significant amounts of time due to the harsh environment of the wastewater flow. A sensor that is immersed in the wastewater can get obstructed by solids in the flow, which compromises its data. Thus, wastewater quality events may not be detected promptly.
In contrast, A-EYE contactless sensors do not touch the wastewater flow. They are not subject to interference or obstruction from the flow and can provide a continuous stream of data to the data loggers, which transmit it to the software component.
Automatic sampling
When necessary, such as when a wastewater event is detected, the software component automatically triggers sampling. The sample is then sent to the lab for analysis. The lab sample analysis provides more detailed information about the wastewater's chemical composition.
Figure 10 Automatic sampler & lab results as shown in the dashboard
Continuous monitoring, coupled with automated sampling triggered by events detected by the Kando software component, eliminates the need for infrequent and labor-intensive grab and composite sampling. The lab results can then be used by the utility as a transparent communication tool to initiate a dialogue with the discharger.
Detecting wastewater quality events automatically
When the Kando Software component automatically detects a wastewater event, it alerts the user in time to take the necessary steps to prevent the WWTP from compromising its efficiency or shutting down.
The Kando software component detects wastewater quality events based on data alone, eliminating the need for human input or graph analysis, as well as eliminating the potential for human error.
To detect wastewater quality events, the Kando software component uses Artificial Intelligence (AI) and Machine Learning (ML) technologies to compare incoming data from the A-EYE sensors with a large amount of data from various sources. When the Kando Software component detects abnormalities in the wastewater quality, it flags this potential wastewater event and alerts the user.
Shifting the focus from detection to prevention
AI- and ML-based technologies look for patterns in large quantities of data. Their power lies in the amount of data they can process, which is far above and beyond what any human, or even a team of humans, is capable of.
Sources of data
In addition to the measurements obtained in real-time by the A-EYE sensors, the kando software component uses the following sources of data for its AI/ML component:
Proprietary historical data
Sector information
Publicly available data
Weather data
Permits
Violations
Online data
Geographical data
Lab sampling results
Contact sensor data
Contactless sensor data
Figure 11 Data flow hierarchy to showcase the order of data sources used by the software component
Observing changes in wastewater quality over time
The Kando software component includes a dashboard that provides an overview of the entire wastewater network, displaying pollution levels in each area based on data from our AI/ML-powered software. It lists recent wastewater quality events and their sources.
The dashboard presents a visual representation of the data the system collects over time. The user is able to see, at a glance, how wastewater quality changes over time.
Figure 12 The Kando dashboard
The dashboard allows the user to view data relevant to a specific area.
Figure 13 View of a specific area
The user can also view data relevant to a particular Significant Industrial User (SIU).
Figure 14 View of a specific SIU
What happens when a wastewater event is detected
When a wastewater event is detected, the Kando software component uses geographical information (2nd party data) to identify the event’s source and calculate the time until the polluted wastewater reaches the WWTP.
Figure 15 Event overview displayed by the Kando software component
Armed with this information, wastewater managers can respond to wastewater quality events on time and take the appropriate actions to change network behavior.
Deploying A-EYE sensors
Kando’s software attribute, the proprietary algorithm, determines the optimal locations for each of the A-EYE sensors in the network, based on detailed analysis of the area’s Significant Industrial Users (SIUs) combined with domestic sanitary dilution.
Figure 16 Technician installing the hardware components in a manhole
Existing customers: Switching from traditional sensors to A-EYE sensors
The process of transition from traditional sensors to A-EYE sensors happens behind the scenes, invisible to the user. There is no change on the Kando software attribute or the dashboard.
The difference lies in the more accurate wastewater event detection that users experience after the transition. By replacing traditional sensors with A-EYE sensors, Kando replaces the sensor that provides real-time data to support the AI/ML software component that identifies wastewater events.
When A-EYE sensors are installed, traditional sensors are left in place for a defined time period, until the A-EYE sensors are calibrated. After that initial phase, traditional sensors are removed.
The following is a graph depicting the software's performance over time during the transition period:
Figure 17 Graph comparison of wastewater event detection based on A-EYE sensors vs. traditional sensors
The curve that begins as a flat line represents wastewater event detection based on the A-EYE sensors. The second curve represents wastewater event detection based on traditional sensors.
As can be seen from the graph, in the beginning, data from A-EYE sensors alone were calibrated within a few hours.
New customers: A-EYE sensors
For new clients, Kando exclusively installs A-EYE sensors
What training is offered by Kando to adopt the system
When you incorporate Kando’s wastewater intelligence solution into your wastewater management, Kando provides you with the training services you need to benefit from the wastewater intelligence solution insights into your network.
During the initial month of using the Kando Wastewater Intelligence Solution, our hardware component collects real-time data, while the software component processes it and the Kando services component supports the utility team in defining priority areas and identifying potential contributors to wastewater quality events in collaboration with their wastewater experts and data scientists and strategize accordingly.
What service is available from Kando
To further enhance your experience, Kando offers comprehensive support services in the form of hardware deployment and maintenance, access to skilled data scientists who continuously update the software and expert wastewater professionals. These services ensure that your wastewater intelligence is always optimized and up-to-date.
When a consistent source of pollution is identified, Kando helps utilities interpret the data and communicate with the contributor, with the goal of reducing pollution and improving wastewater quality and therefore operations of WWTP.
Summary
Kando’s Wastewater Intelligence is an invaluable resource to any wastewater management network. It continuously provides an understanding of the wastewater quality in the network, automatically detects events, and alerts the user in real-time. With the aid of wastewater intelligence provided by Kando, users can maximize their WWTP’s efficiency, improve wastewater quality inflow, prevent shutdowns, improve infrastructure lifespan, and ensure transparency on wastewater activity internally and externally. Kando’s wastewater intelligence is especially effective with the use of Kando’s contactless A-EYE sensors.