The Forrester Wave™: AI/ML Platforms, Q3 2022, released on July 12, not only ranks C3 AI as a Leader, but, according to the report, we “…could become the de facto AI platform for the world’s most complex industries.”
The report is a useful resource for organizations looking to deploy AI/ML at enterprise scale. Readers can request a complimentary copy here.
Of the 15 vendors in the Forrester report, C3 AI received the top score in the Strategy category, and the highest scores possible for the following criteria:
- Product Vision
- Application Tools
- Application Accelerators
- Market Approach
- Architecture Run-time
- Architecture Security
- Data Features
- Partner Ecosystem
We’ve invested $1 billion over the past dozen-plus years, and today we deliver the only comprehensive, fully integrated AI application development platform and portfolio of prebuilt enterprise AI applications. Our products power the world’s largest enterprise AI deployments at organizations including Shell, Koch Industries, the U.S. Department of Defense, Bank of America, LyondellBasell, Philips, and ENGIE.
“At a time when most vendors focus on tools for data scientists, C3 AI has always envisioned a platform approach to AI,” Forrester analysts and report authors Mike Gualtieri and Rowan Curran wrote. “Ahead of its time, C3 AI’s strategy is to make AI application-centric by building a growing library of industry solutions, forging deep industry partnerships, running in every cloud, and facilitating extreme reuse through common data models.”
Let me highlight a few key factors underlying for us this recognition of C3 AI as a Leader in AI/ML platforms.
AI/ML Architecture Supporting Entire Range of Users
Our focus has always been on enabling enterprises to rapidly achieve high value outcomes. Forrester gave C3 AI the highest possible score (5) in the architecture security and runtime criteria. We are the only company that provides end-to-end metadata abstraction (through the C3 Type System) and no/low/deep-code tools to simplify all aspects of enterprise AI application deployment — data integration, persistence, enterprise services, processing, app logic, workflows, ML pipelines, and UI. Our tools support the entire ML model development life cycle and every type of user: developers, data scientists, analysts, and business users.
Deep code tools include a VS Code extension (with autocompletion, debugging, documentation, metadata deployment), a JupyterLab interface with the C3 Type System SDK, distributed in-memory compute, and auto-metadata generation. No-code/low-code tools include a visual canvas to integrate data, manage data models, configure AI/ML, and build, extend, or integrate AI applications. No-code tools provide visual drag and drop capabilities for citizen data scientists or business analysts to explore data, build models, and prototype applications for specific use cases.
All these tools are interoperable and accelerate application development and production operations. All model or application artifacts integrate with third-party applications.
Comprehensive Data and Feature Engineering Capabilities
We provide automatic feature extraction and down-selection. All features are described as metadata transformations and managed in the C3 AI Feature Store, and lineage from source to features is tracked. All data corresponding to features can be materialized, stored, and versioned.
Feature transformations (e.g., SVD, PCA) can be defined as ML pipeline steps using custom programming and open-source libraries (e.g., Scikit-learn, NumPy, Pandas). C3 AI offers a no-code capability to develop and manage features, which can be packaged for use in production applications.
Extensive ML Training Capabilities
Ease and comprehensiveness of ML model development are important platform requirements, and C3 AI excels in both aspects. We provide 100+ differentiated algorithms for a range of 40+ enterprise use cases, including for spatial and temporal modeling, probabilistic forecasting, digital twins, MILP models, stochastic optimization, recommenders, and deep learning auto-encoders. Data scientists at Koch Industries, for example, are using over 20 C3 AI differentiated algorithms to address use cases in reliability, demand forecasting, optimization, and supply chain.
C3 AI integrates with all open-source and third-party frameworks. Open-source libraries and frameworks include Spark MLlib, TensorFlow, Keras, Scikit-learn, cuDNN, SciPy, Caffe, PyTorch, AWS ML, Lex, Polly, Rekognition, Azure ML, H2O.ai, Stanford Core NLP, NLTK, spaCy, fbProphet, StatsModels, and XGBoost. Customers can install any library or custom packages.
We provide model templates for common business problems (e.g., fraud, supply chain disruptions). And we provide an innovative deployment framework to quickly deploy models per segment or entity. C3 AI automates data lineage, feature tracking, and MLOps tasks, including versioning, scale out, performance, re-training, and hot swapping.
C3 AI is uniquely focused on the transition of AI/ML models from prototyping to production operations at scale. The C3 AI Platform enables enterprise data science teams to prototype their algorithms against production data, and then directly attach those algorithms to application object models for production operations at scale. The C3 AI Platform removes the need for additional work, or production AI/ML model refactoring, by enterprise ML or application engineers.
The C3 AI Platform simplifies and streamlines the development and operations of different types of models and pipelines (e.g., ML, Deep Learning, optimization, physics-based models). C3 AI ML Pipelines support complex, multi-step algorithm graphs – with individual steps that may use different techniques (e.g., ML, Deep Learning), runtimes/hardware profiles, and model libraries.
These features all accelerate the transition of algorithms from “prototyping to production” and is a significant reason why our customers choose our technology.
Another factor we believe contributed to C3 AI’s recognition as a Leader is our growing portfolio of prebuilt AI applications that significantly accelerate customer value. Forrester gave C3 AI the highest score available for the application accelerator criterion. We provide not only prebuilt data models, ML models, and workflows, but also complete applications that are customizable and extensible and can be supplemented using common data models and methods with custom-built applications for other adjacent use cases.
Please note that the scope of this Forrester Wave is focused on AI and ML Platforms. A significant part of C3 AI’s products and business is outside our AI and ML platform product and comprised of rich, industry specific, workflow-driven, enterprise AI applications (e.g., CRM, Supply Chain, Reliability, Financial Services, Public Sector, Oil and Gas).
Today we offer more than 40 of these turnkey enterprise AI applications, the most extensive portfolio of prebuilt enterprise AI solutions on the market. We remain very focused on these prebuilt AI applications – a major differentiator from other platform vendors that significantly contributes to C3 AI’s market leadership – and customers select us for the strength of our AI applications as much as the strength of our platform.
C3 AI Applications are fully extensible through metadata with end-to-end functionality: integrations, ontologies, object models, features, AI/ML, workflows, and UI. We also partner closely with Microsoft and Google Cloud to jointly develop solutions targeted to specific industry verticals. Last month, for example, we launched a suite of supply chain applications developed in partnership with Google Cloud.
Rich Data Visualizations
C3 AI provides a comprehensive range of data visualization tools and capabilities for users of all types. For data analysts, C3 AI Data Vision is the most robust tool available today that accelerates the exploration and visualization of data as graph networks, with node and link analysis, geo-spatial, and time-series charts.
For citizen scientists, we provide a no-code tool to accelerate data exploration with scores of pre-packaged column operators, filter and split functions, math and statistical operators, object and text operators, graph functions, and chart components and configurable dashboards.
And for advanced data scientists, we provide a deep-code Python interface in a JupyterLab front end, as well as distributed in-memory data processing capabilities to accelerate advanced data exploration, wrangling, transformation, and visualization using open-source Python tools (e.g, Matplotlib, Plotly, Seaborn).
Robust Security and Multi-Cloud Runtime
Strong security and the ability to support applications running on any infrastructure are key requirements for an enterprise AI/ML platform, and Forrester awarded C3 AI top scores for both security and multi-cloud portability. Our products support the most security-sensitive, mission-critical operations in diverse industries including defense and intelligence, financial services, energy, and industrial manufacturing. And we’ve received virtually every significant security certification.
C3 AI is infrastructure-agnostic and fully containerized (Kubernetes-based), enabling the runtime to be deployed on public clouds (AWS, Azure, Google Cloud), private clouds, on-premises, or the edge. The decoupling of the application logic from infrastructure services enables C3 AI to use any cloud service and easily swap individual services as improved capabilities become available without having to re-write code. This capability protects the customer’s technology investments and prevents cloud lock-in, a growing customer concern.
Industry-Proven, Customer-Driven Products and Technology
Forrester awarded C3 AI the highest scores possible (5) on multiple criteria of strategy, including product vision, market approach, performance, and partner ecosystem. Customers highly value the fact that our products and technology have been tried, tested, and proven in the most demanding production environments in complex industries. Today, customers are deploying C3 AI products at massive scale, generating more than 2 billion predictions daily.
Our product vision maps very closely to the current and future needs of our customers and is informed by continuous customer input and learnings from customer deployments. We conduct quarterly customer satisfaction surveys and meet monthly with customers regarding the product roadmap. We review all community posts and perform quarterly customer advisory board meetings to ensure strong alignment with our customer base. And then we publish a six-month roadmap and statement of direction to our customers based on this feedback.
Our latest release (v8) was based on inputs from 3,500 customer meetings, 1,500 training feedback reports, 16,500+ community posts, 550 CSAT surveys, 12,000 Zendesk tickets, and 500 product roadmap sessions.
Powerful Partner Ecosystem
The final dimension I want to highlight here is our partner ecosystem, for which Forrester awarded C3 AI the highest score possible (5). We have established a broad range of global strategic partners, technology partners, and integration partners. We conduct joint product development and go-to-market activities with global strategic partners like Microsoft, Google Cloud, AWS, Baker Hughes, Raytheon, PwC, EY, Accenture, ENGIE, and FIS.
Importantly, C3 AI is seen as highly complementary to many AI point solutions in order to achieve enterprise value. For instance, Google Cloud, AWS, and Microsoft work closely with C3 AI to integrate Vertex, SageMaker, and Azure ML respectively with C3 AI to ensure scalable AI applications and business value for end customers. Shell, for example, uses Databricks for data ingestion and preparation and C3AI for AI/ML and enterprise scalability.
All these technology relationships are strategic and supported at the development GM level and CEO level. Moreover, our open API framework allows customers to plug in other tools, enabling us to continue to expand our partner network.
While we are honored to be named a Leader in The Forrester Wave: AI/ML Platforms, Q3 2022, we owe our greatest thanks to our customers. These organizations are among the most sophisticated and demanding enterprises operating in the world’s most complex industries — and they push us every day to deliver more and solve previously unsolvable problems. If your organization is looking to solve high-value, difficult problems with AI/ML, we invite you to get in touch with us.