Caylent Catalysts™
Generative AI Strategy
Accelerate your generative AI initiatives with ideation sessions for use case prioritization, foundation model selection, and an assessment of your data landscape and organizational readiness.
Symmons Transforms Water Management & Conservation with AI and Machine Learning
Generative AI
Artificial Intelligence & MLOps
Solution Implemented
→ AI-powered anomaly detection identifies abnormal water usage patterns, leaks, and inefficiencies in real-time.
→ ML-driven predictive maintenance detects potential failures before they escalate, reducing downtime, emergency repairs, and overall maintenance costs.
→ Scalable MLOps infrastructure ensures continuous model retraining, optimizing detection precision over time.
→ Generative AI contextualizes insights, providing automated troubleshooting recommendations based on historical data and maintenance manuals, minimizing reliance on specialized personnel.
→ Cloud-based architecture leveraging Amazon SageMaker AI, Bedrock, DynamoDB, Lambda, API Gateway, and OpenSearch ensures scalable, high-performance data processing and real-time AI model deployment.
Outcomes Expected
→ Projected to save 5 billion gallons of water over the next five years, directly supporting ESG, sustainability, and LEED-aligned environmental goals.
→ Currently monitoring over 1 billion gallons of water annually, already saving over 80 million gallons in 2024 alone across 74 high-usage sites.
→ Identified and resolved 400+ leak incidents in the past year, preventing an estimated $5 million in water damage.
→ Hotels and multifamily facilities experienced 8–12% reductions in water heating costs by adjusting temperatures based on AI guidance.
→ Enabling faster response times, reducing mean time to resolution (MTTR) by over an hour per incident compared to traditional methods & saving $10k+ per incident through reduced emergency labor, contractor costs, and guest reimbursements.
→ Proactive maintenance at scale, with predictive recommendations across hundreds of properties, each proactive action saving an average of $10k+ by preventing full system failures.
→ Established the foundation for self-healing building infrastructure, enabling future autonomous, AI-guided system adjustments, further increasing operational efficiency and decreasing human intervention.
Transforming Water Management
Reliable water systems are easy to take for granted—until a malfunction leaves businesses scrambling to fix costly damage and disruptions. Leaks and undetected plumbing failures waste millions of gallons of water each year, driving up expenses and harming the environment. These issues can escalate over time, damaging critical infrastructure and resulting in repair costs that can scale up to millions of dollars.
Traditional Building Management Systems (BMS) have long been the gold standard for monitoring building operations, but their limitations are becoming increasingly apparent. Not only do these systems come with prohibitive costs ($500K+ upfront), but they also require specialized engineers to manually interpret data—a process that is both time-consuming and inefficient. As a result, most facilities lack real-time anomaly detection, relying instead on reactive maintenance that allows leaks, inefficiencies, and system failures to persist unnoticed.
Symmons is changing this dynamic by making cutting-edge water management solutions accessible to a broader market. By leveraging IoT and AI-driven automation, Symmons can bring affordable, scalable, and easy-to-deploy water conservation tools to hospitality, commercial real estate, and multifamily housing markets.
Symmons: A Legacy of Innovation
For over 85 years, Symmons has been at the forefront of water management innovation, blending precision engineering with cutting-edge technology. Founded in 1939 by Paul Symmons, the company revolutionized the plumbing industry with the invention of the pressure-balancing shower valve, a groundbreaking design that remains the industry benchmark.
Today, Symmons continues to push the boundaries of innovation, integrating technology to optimize water conservation across multifamily housing, hospitality, and commercial real estate. With 220+ employees and $100M+ in revenue, the company manufactures its products in the U.S. and serves global brands like Bellagio, Marriott, Hilton, and Simon Malls.
In 2019, it entered the IoT space with Evolution®, a smart water management system that delivers real-time insights into water usage, leaks, and temperature, enabling proactive facility management. Recognizing an opportunity to enhance insights and accelerate response times, the company partnered with Caylent to integrate AI-powered automation and Generative AI capabilities into its platform.
Turning Data into Action with ML & Generative AI
The Evolution® Building Management System provides granular insights into water usage trends. However, for many customers, this translates into millions of data points streaming in from hundreds or even thousands of properties. Without the right tools, interpreting this vast amount of information and turning it into meaningful actions can be time-consuming and complex. Left unaddressed, inefficiencies and leaks persist, compounding over time and driving up costs. Transforming this data into clear, actionable insights is key to maximizing the impact of Symmons' platform.
To bridge this gap, Symmons partnered with Caylent and AWS to integrate ML-driven anomaly detection and Generative AI-powered insights, making real-time water monitoring more intelligent and proactive. By leveraging AI models trained on historical trends and system documentation, Symmons' platform can pinpoint anomalous water usage patterns and suggest tailored solutions to facility managers in real-time.
By continuously learning from live data, the system improves detection accuracy over time, ensuring ongoing enhancements in water conservation and operational efficiency. With this AI-powered approach, Symmons is transforming raw data into immediate action, helping businesses save water, reduce costs, and operate more sustainably.
Driving Significant Cost Savings, Efficiency & Sustainability
The Evolution® Building Management System has already demonstrated significant impact, helping commercial properties and hotels detect inefficiencies, reduce water waste, and lower costs through real-time IoT monitoring. In 2024 alone, the platform helped customers save over 80 million gallons of water across just 74 high-usage sites—part of a larger effort to monitor over 1 billion gallons of water annually across the portfolio. Symmons has set a bold target of saving customers 5 billion gallons of water over the next five years.
Massive Cost Savings Through Early Detection
Across the board, Evolution has delivered tangible cost savings by detecting inefficiencies, leaks, and usage anomalies early, before they escalate into costly problems. One hotel saved $200,000 annually by identifying continuous cooling tower overflow, while another reduced insurance premiums by $50,000 after implementing leak detection safeguards during construction. These types of interventions aren’t outliers: customers have saved thousands to tens of thousands of dollars per year by catching issues like stuck irrigation valves, cooling tower drift, and underground leaks that would otherwise go unnoticed. At a major retail center, a broken float switch alone was wasting nearly 3 million gallons per year and has now been fully remediated.
Faster Resolution, Fewer Service Disruptions
The system has also proven effective in improving response times and operational efficiency. By providing real-time insights and prescriptive recommendations, Evolution has reduced mean time to resolution by over an hour in hospitality use cases, giving facility engineers time back and preventing service disruptions.
Predictive Maintenance That Pays Off
Evolution is already helping customers take proactive action across hundreds of properties. On average, each incident prevented upwards of $10,000 in guest reimbursements, emergency labor, and brand damage. The platform has identified and helped resolve over 400 leak incidents and flagged 500+ sustained scalding conditions—two areas that directly impact building integrity and occupant safety.
Smarter Energy Management and Heating Efficiency
Beyond leak detection, Evolution supports sustainable energy management through intelligent water heating optimization. Hotels and multifamily buildings have reduced hot water heating costs by 8–12% simply by adjusting temperature setpoints based on AI recommendations. Another high-rise building cut overnight reheat coil temperatures to save $5,000 per year in energy waste. These insights not only reduce utility bills but also help customers meet energy conservation and carbon reduction goals.
Evolving Toward Self-Healing Infrastructure
Looking ahead, Symmons has ambitious goals for the Evolution® platform. In the near term, automated anomaly detection will become even more precise, with self-learning models reducing false alerts and prioritizing critical issues. AI-powered data contextualization will help facility teams understand the root causes of anomalies faster, while predictive maintenance models will forecast potential infrastructure failures—allowing teams to take preventative action before issues escalate.
Longer term, future enhancements will enable real-time system adjustments, where AI dynamically manages systems to maximize efficiency without human intervention. As these models advance, Evolution® is progressing toward true self-healing infrastructure: systems that detect, diagnose, and autonomously correct inefficiencies in real time. The result? Smarter, more resilient buildings with minimal operational overhead.
Technical Challenge
Symmons needed an ML-powered anomaly detection system that could automatically detect, classify, and contextualize water inefficiencies while integrating into its existing Evolution® platform. A robust MLOps framework was also necessary to continuously retrain models, deploy updates seamlessly, and scale across thousands of properties. Additionally, Generative AI-powered insights were required to help non-expert facility managers quickly diagnose issues, receive prescriptive recommendations, and take corrective actions without needing deep technical expertise. A low-latency cloud-architecture leveraging AWS services was key for real-time data processing, searchability, and AI-driven decision-making at scale.
Solution
Caylent designed a scalable, flexible, and future-proof anomaly detection and AI-powered monitoring system tailored to Symmons’ needs. The solution integrated IoT sensors, AI-powered analytics, and cloud automation, making real-time monitoring accessible, actionable, and continuously improving over time.
AI-Powered Anomaly Detection
The foundation of the solution is a machine learning-based anomaly detection system designed to identify abnormal water usage patterns, leaks, and inefficiencies with minimal false positives. This involved:
- Time-Series ML Models: Trained on historical sensor data (flow rates, temperature fluctuations, humidity levels) to recognize deviations from normal operating patterns.
- Adaptive Learning: Self-improving models that adjust anomaly thresholds dynamically based on seasonal patterns, property type, and occupancy trends.
- Multi-Sensor Correlation: AI processes data from multiple IoT sensors per facility, allowing for cross-referenced anomaly validation to improve accuracy.
- Event Classification: AI categorizes detected anomalies (e.g., persistent leaks, irregular consumption, equipment malfunctions) and prioritizes urgent alerts for faster resolution.
Scalable MLOps Infrastructure
To ensure continuous model improvements, versioning, and deployment at scale, an MLOps pipeline was built using AWS SageMaker, Lambda, DynamoDB, and API Gateway. This enables:
- Automated Model Training & Retraining: ML models continuously improve using live operational data, ensuring higher detection precision over time.
- Real-Time Inference at Scale: AI inference services run on AWS Lambda and SageMaker endpoints, providing sub-second anomaly detection.
- Version-Controlled Deployment Automation: A fully managed pipeline ensures validated, tested AI models are deployed seamlessly, minimizing disruptions.
- Edge & Cloud Hybrid Processing: AI models are optimized for low-latency edge computing, enabling faster local anomaly detection while aggregating broader insights through cloud AI processing.
- Experimentation-Ready AI Model Management: Symmons can test different AI models dynamically, compare outputs, and switch to the best-performing model without redeploying the entire system.
- Flexible AI Knowledge Base Architecture: AI can integrate different vector databases, allowing teams to test different knowledge bases and refine AI accuracy dynamically.
Generative AI Co-Pilot for Insights & Prescriptive Actions
One of the most transformative features of the solution is Generative AI-driven contextualization, which bridges the gap between data insights and human decision-making. This is achieved through:
- Retrieval-Augmented Generation (RAG): AI models trained on technical manuals, maintenance logs, and past anomalies, allowing facility managers to receive specific, explainable recommendations based on their property’s unique conditions.
- Automated Troubleshooting Suggestions: Instead of simply flagging anomalies, AI suggests corrective actions, such as adjusting recirculation pump flow rates, flushing clogged risers, or recalibrating valve pressure settings.
- Conversational AI Interface: Facility managers can ask AI-powered assistants for real-time troubleshooting help, allowing even non-expert users to understand and act on insights.
- Historical Anomaly Contextualization: AI tracks recurring issues over time, offering insights into long-term infrastructure inefficiencies that may require capital improvements or proactive maintenance.
Real-Time Data Visualization & Facility Dashboard
To bridge AI-powered insights with human decision-making, the Evolution® facility dashboard provides real-time data visualization and alerts. Facility teams receive instant notifications via SMS, email, and mobile apps, ensuring critical anomalies are addressed immediately. Interactive dashboards allow managers to analyze efficiency trends, track cost savings, and benchmark performance across multiple properties. The AI also generates predictive maintenance schedules, allowing teams to plan repairs proactively rather than responding to failures.
Cloud-Native & Scalable Infrastructure
The system was built using AWS-native services that allow for real-time streaming, high-availability AI processing, and enterprise-grade security:
- AWS Bedrock for Generative AI & RAG: Enhances AI-driven insights by leveraging Retrieval-Augmented Generation (RAG), allowing the system to contextualize anomaly detection data with water system theory, building metadata, facility-specific water usage trends, and issue history.
- Amazon SageMaker AI: Enabled model training on historical sensor data and automated retraining using live operational data, ensuring improved detection precision over time.
- AWS IoT Core: Collects and processes millions of real-time sensor readings from across facilities.
- AWS OpenSearch as VectorDB: Enables fast search and retrieval of historical anomaly records and AI-generated recommendations.
- AWS DynamoDB: Serves as a highly scalable, low-latency data store for structured AI model metadata.
- AWS API Gateway & Lambda Functions: Ensure real-time communication between AI models, IoT sensors, and facility management tools.
