The Language of Business with Jim Snabe | C3 Transform 2026

Well, thank you for setting expectations low. I want to congratulate Stephen and the C3 team for transforming C3 at a time when AI will transform business with value, not just tokens. I was reminded of the big change in technology by a conversation with my daughter. When she was five, she asked what I do at work. I described meetings and conferences, and she said, 'Don't you do any real work?' Now she's 27, and when I asked her the same question, she said, 'No, but my agents do.' Despite that, the first word that came to mind for this keynote was chaos. We live in a world of geopolitical chaos and technology chaos—a race for power and a race for intelligence. Everyone is trying to build AGI, but there's lots of promises and little value. We're running out of energy and, more importantly, running out of words to train these models. They're trained on all the words ever written, and now we're making synthetic language—AI training AI. It reminds me of mad cow disease. I hope our LLMs don't get word disease. Many AI companies spend obscene money trying to make the best LLM. But the amount of data a large language model is trained on is roughly what a five-year-old sees in their first five years. Would you run your company with a five-year-old? I asked ChatGPT how to find the fastest of six horses, and it suggested three races. My five-year-old daughter would have just run them all together. That's where we are. These models are trained on all the lies of the internet, so of course they hallucinate. They predict the next word; in business we predict the next event and make decisions. We need to focus on the language of business: value, prediction, and scale. Value in AI is currently in the infrastructure stack, but it will shift to applications, just like electricity. Prediction in business means predicting failures, not the next token. Scale in business is decentralized, not centralized data centers. We need to move from AI that describes the world to AI that acts in the world—agentic AI. But imagine a five-year-old English teacher taking action in your company—that's scary. We need design for autonomy, not extending lane assist. This requires an enterprise platform with unified data, orchestration, and governance. My three recommendations: start with the business value, train your own AI on your high-quality data, and standardize on one real AI platform. Master it, don't build it. When it comes to choosing technology in an enterprise, I suggest dictatorship. Will AI ever be as smart as humans? Yes, but it will be a combination of horizontal language models and domain-specific models trained on your data. The key is the orchestration layer. AI is a tool—an amplifier of human intent. Leadership is about dreams and details. This is a leadership moment. When someone asks what you do, tell them you are shaping the future of agentic AI in a responsible way. Thank you.