The last 12 months have been exciting for C3 IoT. We reported record financial results last week. But what really excites me is the progress our customers have made to capture value from their AI-based digital transformations using the C3 IoT Platform and applications. Our product, engineering, data science, and professional services accomplishments have helped our customers to rapidly deliver new AI applications – at scale and faster than ever before, using the C3 IoT Platform.
Here is a sampling of some of this year’s accomplishments that I am personally very excited about:
Real-Time AI and IoT
The C3 IoT Platform has enhanced support for ingestion and processing of high-frequency time series data (e.g., second or sub-second). We have made significant improvements in how we store, index, and access data in file systems such as Amazon S3 and Azure Blob Store – while also supporting time-series based analytics on highly responsive infrastructure.
Across all deployments, the C3 IoT Platform now consumes and analyzes data in production from more than 300 million devices. Relevant deployments include:
- Ingesting and analyzing telemetry across 400,000 heavy equipment assets, streaming data through Azure EventHub to the C3 IoT Platform, resulting in over 24 petabytes of data in the first year.
- Ingesting 1-second data from 100,000 sensors to assess gas well torque-up issues in near-real time.
- Processing 1 sec - 5 min interval data from 1.5 million devices to optimize solar production.
- Processing data from more than 20 million devices—with new readings every 1-15 minutes, normalized against enterprise, operational, and external data to make 12 million predictions every day for predictive maintenance, customer segmentation, and fraud detection.
The C3 IoT Platform also introduced data virtualization support enabling customers leverage their existing investments in data lakes, warehouses, or legacy enterprise systems and eliminating the need to duplicate data. C3 IoT customers can model data in existing systems as “External Types” and leverage core platform functionality such as data persistence, time series modeling, normalizing and machine learning.
Multi-Cloud and Hybrid
A major initiative in FY ’18 was the development and delivery of a multi-cloud strategy. C3 IoT made significant investments in Docker and Kubernetes support for deployments, enabling the C3 IoT Platform to run on public or private cloud infrastructure. In April, C3 IoT announced a strategic partnership with Microsoft to deliver our low-code, high-productivity PaaS on Microsoft Azure. As part of the close collaboration with Microsoft, C3 IoT integrated 23 new Azure services in FY ’18, enabling customers to leverage best of breed cloud services across AWS or Azure (e.g., Azure IoT Hub, AWS IoT, AWS Kinesis, Azure EventHub, and Azure ML). C3 IoT also announced its partnership with Intel to deliver the C3 IoT AI Appliance powered by Intel AI – enabling organizations with stringent data residency needs to design and deploy AI and IoT solutions on the C3 IoT Platform.
We released a new version of C3 Tools, providing a low code or no code development interface that encourages developer collaboration and enables users to configure data integrations, data models, analytics, AI / machine learning, user interfaces and manage APIs. C3 Tools provides an integrated experience to deploy and operate the developed applications, and includes continuous integration, automated testing, and processes to manage and promote code across environments.
C3 Ex Machina enhancements
C3 Ex Machina introduced several data preparation, machine learning, and data visualizations enhancements in FY ’18. C3 Ex Machina is now in production with several customers, helping their business analysts tackle complex challenges such as customer segmentation, fraud detection, pricing, and rate analysis use cases. On the backend, C3 Ex Machina now supports Spark 2.2.
Machine Learning pipeline and model management
Techniques like deep learning and natural language processing are making machine learning algorithms more complex. As a result, designing and managing analytic pipelines and models that are human interpretable, reproducible, and testable is more of a challenge – particularly for enterprise scale models that make ongoing predictions as data sources update and data streams change.
To solve this challenge, the machine learning C3 Pipeline APIs have been enhanced to let teams write their own custom R and Python code, or link together multiple pipelines, each with their own frameworks, for more complex use cases. Pipelines are stored, access controlled, and versioned centrally, and model management tools are part of a deployed application package.
Our Next Generation UI
C3 IoT’s Design Innovation Team launched a major initiative to develop our Next-Generation UI Framework that includes a standardized design language, cutting edge data visualization tools, an easy to use design environment called C3 Application Studio™, and a complete set of documentation for developers. Design and implementation of the new UI framework is well underway, with new components and capabilities planned throughout the year that will make it easier than ever to design and configure powerful IoT application user interfaces using our platform.
Applications - Inventory Optimization
A large U.S. manufacturer deployed C3 Inventory Optimization at scale earlier this year to optimize their inventory levels while maintaining confidence in meeting customer service. The application—which provides inventory recommendations in near real-time—has been well received and is being expanded to additional factories in the U.S. and Europe. C3 Inventory Optimization combines the flexibility of the C3 Type System, embedded AI algorithms, stochastic optimization, and a scale-out processing architecture to enable inventory levels to be managed dynamically at the individual SKU and factory-levels.
A large global aircraft equipment manufacturer also successfully completed an initial project at two factories - using C3 Inventory Optimization to predict supplier delays or de-commits. As a result, the manufacturer can avoid significant costs related to stranded inventory and is able to improve service levels. The manufacturer is now developing a rollout plan across additional manufacturing facilities.
Applications - Anti-Money Laundering
Regulatory and organizational challenges are driving financial services compliance teams to change how they identify money laundering and terrorism financing activities. This year, we launched a new fraud detection application for financial services and deployed it at a top 50 global bank.
C3 Anti-Money Laundering’s algorithms identify suspicious activity issues with much higher precision than conventional methods by using big data sets to form a continuously evolving and holistic view of a customer’s risk. C3 Anti-Money Laundering supports automated closed-loop feedback to improve predictions and update risk profiling.
Applications - Predictive Maintenance
This year, we expanded our AI-based predictive maintenance application to address high value business cases in new industries. Customers can choose from a comprehensive set of industry-specific packages that can be deployed on a base package to deliver out-of-the-box AI-enabled predictive maintenance solutions, at scale, in production, in 6 months or less. New industry deployments include:
- Aerospace and Defense: Over the past 7 months, C3 Predictive Maintenance has been configured for the U.S. Air Force’s E-3 Sentry and C-5 Galaxy fleets. The application’s classifiers and MTBF analytics provide flight-line maintainers with diagnostic information that can significantly increase aircraft uptime.
- Mining: A global mining company is leveraging C3 Predictive Maintenance to monitor the health of mining assets – an important component of the firm’s Mine of the Future program.
- Heavy Industrials: A large engines manufacturer is using predictive maintenance to assess the health of their global installed base of diesel engines.
New features include a built-in framework for anomaly detection and the ability to efficiently manage time-varying equipment relationships (e.g., ability to track whether individual components or systems have been changed or repaired over time). We also implemented deep learning algorithms for predictive maintenance - to further reduce the amount of time data scientists spend on feature engineering.
Applications - Energy Management
C3 Energy Management continues to see high customer interest. The strongest growth is among utilities as they undergo digital transformations, looking to big data and AI analytics to revamp energy efficiency programs and customer operations. With C3 Energy Management, we are leading the industry to shift an energy efficiency focus from customer portals and paper-based reports to a targeted analytical approach using predictive analytics embedded directly in customer journeys. This is evident in recent customer wins, including expanded partnerships at Eversource Energy, ConEdison, and the New York Power Authority. These three utilities together serve over 10 million customers and use C3 Energy Management to analyze, predict, and deliver targeted energy efficiency insights, recommendations, and services.
Training and Customer Support
Training is crucial to the success of our customers. To better serve them, we developed and launched a beta version of our Coursera training class that teaches developers how to develop on the C3 IoT Platform. Taught by C3 IoT experts, the course provides everything needed to start developing AI and IoT applications.
The combined knowledge of our developer and data science community also provides a direct line of communication with customers and encourages collaboration and knowledge sharing. Our online community for C3 IoT developers and data scientists is handling a large daily volume of questions and responses. Topics included data integration, Platform and data science APIs, and the use of visual tools.
Operations and Security
C3 IoT uses best-practice hosting operations and support processes to ensure the integrity and availability of customer systems. These encompass all aspects of reliable system delivery, including maintenance, continuous integration, and delivery (CI/CD), change management, backup/recovery, and system monitoring for performance and availability.
C3 IoT cybersecurity programs are managed using NIST (National Institute of Standards and Technology) best practices and established IT standards. We perform vulnerability (both DAST and SAST) and penetration testing for every new software version using tools like Fortify and engage external security experts like Hackerone. We protect data with AI-based endpoint protection system, intrusion detection and prevention (IDS/IPS), Data Loss Prevention system (DLP) and enforce data encryption and Multifactor Authentication (MFA). These comprehensive measures have enabled C3 IoT to successfully pass rigorous external audits, receive or renew several cybersecurity certifications such as SOC2 Type II, achieve HIPAA compliance, and comply with GDPR requirements.
All in all, the last fiscal year was a momentous one for C3 IoT—with new customers and applications, and a long list of product improvements and new capabilities—to help our customers achieve digital transformation. Stay tuned because there will be much, much more!