Por que os dados da sua empresa nunca vão estar prontos para IA

Most companies think they're ready for AI. They're not. The data has problems, the infrastructure can't handle it, and governance hasn't been established yet. Daniel Avancini, co-founder of Indicium AI, a partner of Anthropic, OpenAI, and Databricks, spent 6 years trying to bring AI from the lab into large companies and learned this the hard way. In this episode of Product Guru's, we talk about what really hinders AI adoption in the enterprise, what separates a demo project from a product that generates value, and why your company's data will never be perfect, and that's okay. What you'll hear: How Indicium built a marketing agency, saw that it didn't work, and returned to its core data and AI business. What changed in the company after the acquisition of Mesh.AI (UK and Europe). The difference between DataOps, MLOps, and AgentOps. The real risk that Anthropic admitted about the next model. Why data will always have problems, and what to do about it. 00:00 Introduction 00:54 The loneliest moment as a co-founder 02:26 The biggest fear when hiring the first people 03:32 Diluting the focus to grow, what Indicum did 06:35 Mesh.AI acquisition: Brazil-Europe cultural challenges 09:20 What is lost when a startup becomes a large company 10:59 How to take AI from the lab to production 13:16 Building trust with Anthropic, OpenAI and Databricks 16:20 Claude Computer Use and agents: efficiency or risk? 19:54 Will AI replace jobs? 22:53 Can companies' infrastructure handle the speed of AI? 28:52 The truth about data that nobody admits 31:47 Closing and job openings at Indicum 🔗 Indicum AI: https://indicium.ai/ 🔗 LinkedIn Daniel:   / daniel-avancini   🔗 Job openings at Indicum: https://indicium.ai/pt-br/sobre-nos/c... Important messages: 🔗 Build a future-proof career, leading smarter and more efficient products with AI. Access: https://go.pm3.com.br/ProductGurus-AI... 10% discount coupon: PRODUCTGURUS 🔗 Pingback is the simple and fast platform to automate marketing and sales without bureaucracy. Create forms, send emails and WhatsApp messages, segment leads and automate workflows while your team monitors everything in the CRM. Access: https://pbck.link/paulo-chiodi