Product Portfolio Overview

Prepared by Thomas M. Siebel, Chairman and CEO | May 2026

Purpose and Portfolio Overview

This overview gives a clear, non-technical understanding of C3.ai's five core product areas: the foundational platform, the developer studio, the library of industry-specific AI applications, the generative AI suite, and the newest AI coding capability. For each we explain what it does, who invented it, what it took to build, how it is protected, and how it fits with the rest of the portfolio.

The portfolio is best understood as an integrated three-layer stack. The C3 Agentic AI Platform is the industrial-grade engine doing the heavy lifting of data integration, model training, and deployment. C3 AI Studio and C3 Code sit above it as the workbenches engineers and analysts use to build applications. Twenty-plus industry-specific C3 AI Applications run on top. C3 Generative AI is a horizontal layer woven through every product, letting any user converse with enterprise data in plain English.

C3.ai has invested approximately 15 years and more than $3 billion in software development to build this portfolio. At a fully-loaded cost of roughly $375,000 per engineer-year, that implies cumulative effort on the order of 8,000 person-years. The per-product figures below allocate that total; they are informed estimates from the Office of the Chairman, not audited figures.

The Foundation — What Everything Else Is Built On

What It Does

The platform is the core technology on which every C3.ai product runs — a Platform-as-a-Service analogous to the operating system of a smartphone: every application depends on it, but most users never see it directly. Its central innovation is the patented C3 AI Type System, a model-driven architecture. Rather than writing custom code to connect each enterprise data source, a developer describes the business objects they care about — an Asset, an Alert, a Work Order — and the platform generates all the integration plumbing automatically. We measure this to reduce application development time by roughly 25× versus building from scratch.

Key Capabilities

• Data unification across hundreds of enterprise systems and industrial sensors simultaneously.
• AI/ML lifecycle management for training, testing, deploying, monitoring, and retraining models at industrial scale.
• Agentic AI orchestration coordinating multiple autonomous software agents through multi-step workflows.
• Cloud-agnostic deployment on Azure, AWS, GCP, or on-premises FedRAMP-authorized government clouds.
• Enterprise-grade governance with role-based access, audit trails, and model explainability for regulated industries.

Why It Matters to the Business

The platform is C3.ai's primary competitive moat. Every AI application we sell runs on it; every new generative AI capability layers on top of it. In 2025 we launched an OEM program allowing third-party technology partners — initially Fractal, SMX, and Groundswell — to build and sell their own derivative AI applications on the platform, creating a potentially significant ecosystem revenue stream beyond our own product sales.

The Developer Workbench — Where Applications Are Built

What It Does

Studio is the integrated development environment engineers, data scientists, and increasingly business analysts use to design, build, test, and deploy applications on the platform. It hides enormous technical complexity behind a visual interface — think of it as the equivalent of Microsoft Word for writing documents, except here the document is an enterprise AI application.

Key Components

• Visual Studio Code Extension — familiar IDE for professional developers, enriched with C3.ai-specific tooling and autocomplete.
• Application Canvas — low-code drag-and-drop design surface allowing non-engineers to build applications visually.
• C3 Generative AI — turns plain-English requests ("Create an alert when temperature exceeds baseline by 15%") into application logic.
• Data Studio and AutoML — data profiling, cleansing, connection, and automated machine learning to accelerate model development.

Strategic Value

Studio addresses one of enterprise AI's biggest friction points: the chronic shortage of qualified AI engineers. By enabling less-technical staff to build applications, and by embedding a generative AI assistant for professional developers, Studio compresses time from business problem to deployed solution — a key driver of customer satisfaction and renewal rates.

The Product Library — Ready-to-Deploy AI for Specific Business Problems

C3 AI Applications are pre-packaged, industry-specific AI products that customers license and deploy without building AI capabilities from scratch — analogous to buying a professionally designed home rather than building from raw materials. Every application runs on the platform and is maintained in Studio, so platform improvements automatically flow to all applications and new generative AI capabilities can be added rapidly.

Flagship Proof Point: C3 AI Reliability

Reliability is our most widely deployed application. It monitors industrial equipment — gas turbines, aircraft engines, electrical transformers, manufacturing lines, offshore platforms — and predicts failures days or weeks before they occur, giving maintenance teams time to intervene. Generative AI is integrated, so technicians can ask plain-language questions ("Why is compressor C-12 flagged?") and receive grounded, explainable answers.

• Shell: Monitors 14,000+ critical assets globally; 20% reduction in unplanned downtime; 15% reduction in maintenance costs.
• Major U.S. Electric Utility: 48% reduction in transformer failures; $40M+ estimated annual economic value.
• Global Beverage Company: 100% of critical filling-machine failures predicted in advance.

Companion Application: C3 AI Inventory Optimization

Inventory Optimization addresses one of the most expensive problems in manufacturing and retail: holding too much inventory while running out of critical parts. The application uses AI to find the optimal stock level for every item at every location, accounting for demand uncertainty, lead times, and supplier risk. A large electronics manufacturer with 100+ inventory managers gained visibility to 10,000 SKUs in six months; a defense conglomerate using the companion Supply Network Risk application increased late-delivery prediction lead time by nine times; a global agribusiness reduced schedule generation time by 96%.

The Conversational Intelligence Layer — Talk to Your Enterprise Data

What It Does

C3 Generative AI brings ChatGPT-style conversational intelligence to the enterprise, with the critical difference that answers are grounded in the customer's own proprietary data, not just general internet knowledge. A refinery engineer can ask why a pump is behaving erratically and receive a coherent, source-cited answer in seconds, synthesized from sensor readings, work orders, OEM manuals, and inspection reports — a task that previously required hours of expert manual data gathering.

Key Technical Concepts

• LLM-agnostic architecture: Customers can choose or switch among OpenAI, Google, and other models without rebuilding applications.
• Retrieval-Augmented Generation (RAG): C3.ai's proprietary technique that grounds every answer in source documents with citations, preventing hallucination.
• AI Agents (Patented): Autonomous software workers that plan and execute multi-step tasks. The awarded patent protects the orchestration architecture.
• Enterprise-grade guardrails: Role-based access controls ensure users see only authorized data, even through a conversational interface.

C3 Generative AI ships in industry-specific versions for aerospace, defense, financial services, healthcare, intelligence, manufacturing, oil & gas, telecommunications, and utilities, plus function-specific versions for contact centers, financial analysis, human resources, supply chain, and sales.

The AI Application Generator — From Plain English to Production Software

What It Does

C3 Code is the most dramatic leap forward in our product history. A user describes an enterprise AI application in plain English; the system builds a complete, production-ready application automatically — data pipelines, AI models, business logic, security controls, and user interfaces. Work that previously required teams of specialized engineers for weeks or months is orchestrated by multiple AI agents working in parallel, using the customer's own enterprise data, in hours.

How It Works

• The user describes the application needed (e.g., "Analyze my global in-process inventory and show me where my most critical shortages are").
• C3 Code's agents interpret the request against the customer's enterprise data context — ERP, supply chain, logistics feeds.
• Multiple agents work simultaneously: one builds the data model, one configures the ML pipeline, one builds the user interface, one sets up security and access controls.
• Within hours a production-grade application is generated — with live data integration, dashboards, AI scoring, and role-based access — ready for technical review and deployment.

What Makes It Distinctively Enterprise-Grade

C3 Code differs from general-purpose AI coding tools like GitHub Copilot in two ways. First, it starts from pre-built Application Packages encoding deep knowledge of how real enterprises operate in manufacturing, energy, and defense — so generated code reflects industrial workflows, not generic software patterns. Second, the applications it produces are deployed through C3.ai's enterprise pipelines with security, audit trails, and governance built in from the start; the generated code is fully portable and customers are not locked to any specific AI model or cloud provider.

Worked Examples

Supply Chain Intelligence Prompt: "Analyze my global supply chain inputs and provide me a new generative AI application for where my parts are."
Result: A production application resolving part status, ETA,supplier exposure, and delay risk from live logistics data, with grounded answers and role-based access. Agents connected ERP, logistics APIs, freight carriers, and supplier portals automatically. Previously weeks of engineering effort.

Asset Performance Optimization Prompt: "Build an application that monitors equipment health across my fleet, predicts failures, and optimizes maintenance schedules."
Result: Fleet-wide health scoring, hourly failure-probability updates, automated work-order generation, and maintenance optimization across 500+ assets in a governed production workflow — generated without a dedicated data science team.

* MLOC = millions of lines of code. Figures are estimates from the Office of the Chairman, calibrated against the cumulative ~$3 billion invested over 15 years.

How the Products Work Together

• The C3 Agentic AI Platform is the infrastructure layer; all other products run on it and inherit its security, scalability, and data integration.
• C3 AI Studio is the creation layer where engineers and analysts build, customize, and maintain applications.
• C3 Code is the acceleration layer that replaces large amounts of manual work in Studio with automatically generated applications.
• C3 AI Applications are the ready-made solutions layer, encoding C3.ai's accumulated 15 years of domain knowledge.
• C3 Generative AI is the access layer — protected by our awarded AI Agents patent — letting any employee extract value from all the layers below.

C3.ai's intellectual property protection is centered on its model-driven architecture, enterprise AI platform methodology, and most recently its agentic AI capabilities. Key issued patents include:

• US 10,817,530 — enterprise IoT application development platform (granted October 2020); inventors Siebel and Abbo.
• US 10,824,634 — enterprise AI and IoT platform (granted November 2020); inventors Siebel and Abbo.
• AI Agents patent — recently awarded to C3.ai; inventors Siebel and Krishnan. Protects the architecture for orchestrating autonomous AI agents in an enterprise-governed environment; underpins both C3 Generative AI and C3 Code.
• Additional patents in prosecution covering generative-AI methods, agent orchestration, and code generation.