From Plain English to a Deployed Enterprise AI App in Hours with C3 Code | Enterprise AI Spotlight

Enterprise AI is moving from months-long engineering projects to something a domain expert can build in plain English in hours. In this episode of Enterprise AI Spotlight, Shrey Satpathy, Product Manager, Platform, sits down with Amir Delgoshaei, Senior Director, Data Science, for a live, end-to-end walkthrough of C3 Code — C3 AI's natural language application builder that lets domain experts build complex Enterprise AI applications without writing code. Together they build a real application from scratch: a production scheduling and optimization tool for a refinery operator that has to accept crude oil deliveries, decide when tankers dock, blend crude across a limited number of storage tanks, and keep final-product inventory healthy, all while avoiding costly penalties. Starting from a plain-language specification, C3 Code generates the application UI, builds production-ready ETL pipelines and an object model on the C3 Agentic AI Platform, recognizes the problem as a mixed integer linear program, writes and tests the optimizer, and deploys it to production. The conversation covers how C3 Code draws on 15 years of C3 AI's battle-tested building blocks and the C3 AI Type System, why enterprise-grade AI needs far more than a UI prototype (user management, data pipelines, an agent control plane, automated testing, and scale), and how the coding agent verifies its own work against a live backend by catching errors and self-healing until it produces working software. Watch the full walkthrough to see a complete Enterprise AI application built in a matter of hours instead of the weeks or months it would take a team the old way. Learn more about C3 Code: https://c3.ai/c3-code Subscribe to the C3 AI YouTube channel:    / @c3ai   Stay up to date on C3 AI: C3 AI Blog: https://c3.ai/blog/ C3 AI X: https://x.com/C3_AI C3 AI LinkedIn:   / c3-ai   #C3AI #EnterpriseAI #C3Code #AgenticAI 00:00 Introduction to C3 Code 00:41 Welcome to Enterprise AI Spotlight 01:28 The customer challenge: refinery crude oil scheduling 02:27 How the problem is solved today: spreadsheets and silos 03:32 Building the application from a natural language specification 04:25 Designing the interface: what-if scenarios and audit trails 05:22 Encoding subject matter expertise and user personas 06:50 The first agent run: from prompt to implementation plan 08:06 Previewing the generated application 09:37 The real power of C3 Code: built-in platform configuration 12:00 Data fusion: agent-built ETL pipelines and ontology 13:35 Inspecting the object model and reusable building blocks 15:41 From business problem to optimization: MILP and battle-tested primitives 18:37 Verified building blocks vs. writing code on the fly 20:01 Inside the optimizer: solver, type system, and implementation 20:38 How C3 Code verifies code: sandbox and a live backend 21:43 Self-healing: executing code and iterating on errors 24:47 Deploying to production 25:51 Using the deployed app: optimizer and embedded agents 27:14 How long it took vs. the old way 29:14 Closing thoughts