Always evolving Podcast Ep. 5: Is Your Data Actually AI-Ready? w/ Andy Thompson

Every company thinks their data is AI-ready. Almost none of them are (and that is ok!) In this episode of Always evolving, host David Black sits down with Andy Thompson, Manager of Data Engineering at evolv, to talk about what actually happens when you plug messy, siloed, ungoverned data into an AI model — and what it takes to fix it. Andy works on the ground with clients every day. His perspective is honest, practical, and full of the kind of insight you only get from someone who's seen every version of this problem across dozens of organizations. What you'll hear: 🧹 Why "clean data" is almost always a myth — big company, small company, it's almost always a governance mess, siloed data, walled gardens, or "Joe's spreadsheet in accounting that's a don't touch it type of process" 🤖 What happens when you plug that data into an AI model — it hallucinates like crazy, because it doesn't have one version of truth. It has 20 📊 Why data cleanliness is becoming a more important issue in the AI era, not less — because now you're removing human eyes that used to catch errors across the lifecycle 🏊 The two types of AI clients evolv works with: those cautiously experimenting, and those who "jumped in before they knew there was water in the pool" 🎯 How to train clients to actually use AI well — start small, treat it like an intern, give it context, ask focused questions 💰 Advice for the CFO who just got handed the AI bill: don't panic, ask what got done, and measure value holistically — not just tickets closed 🤝 Why AI training isn't just for data engineers — executives need it too 🔀 How to use Claude, Cortex Code, Gemini, and multiple models together to synthesize ROI frameworks and get insights no single model would surface alone This is one of the most practical, no-fluff conversations we've had on Always evolving. If you're anywhere in the AI and data journey — from just starting out to managing the bill — this episode is for you. 📍 Chapters: 0:00 "Your data's AI ready" — and then the hallucinations start 0:27 Meet Andy Thompson, Manager of Data Engineering at evolv 0:47 What evolv actually finds when they get to a client's environment 0:57 The clean data myth — governance mess, silos, and walled gardens 1:25 Joe's spreadsheet: the don't-touch-it data problem 1:45 Plugging bad data into AI: why it hallucinates with 20 versions of truth 2:03 Data cleanliness is more important in the AI era, not less 2:30 How evolv trains clients to be AI-ready 2:37 Two types of clients: cautious experimenters vs. those already swimming 2:53 Start small: treat AI like an intern, give it context 3:07 Focused questions get better answers — revenue from sales ≠ revenue from accounting 3:28 Slow down and think holistically before asking the wrong question 4:02 Advice for the CFO handed the AI bill for the first time 4:08 Don't panic — but ask what got done, not just how many tickets closed 5:04 ROI with AI is nuanced — lean into the nuance 5:33 Building in months what used to take years — that's how you measure it 6:06 AI training isn't just for engineers — the C-suite needs it too 6:24 Using Claude and Cortex Code to calculate ROI frameworks 6:41 Ask multiple models, synthesize the answers, go down the rabbit hole 💬 Ready to get your data AI-ready? Visit https://evolvconsulting.com/contact 👥 About Always evolving Always evolving is evolv Consulting's podcast series featuring real conversations with our practitioners on AI, data, and what it takes to transform a business. Subscribe for new episodes. 👥 About evolv Consulting evolv is an Elite Snowflake Services Partner and AI-first data transformation firm helping enterprises move from data potential to measurable business outcomes — fast. 🔗 Learn more: https://evolvconsulting.com #AlwaysEvolving #AIReadyData #DataEngineering #AITransformation #evolv #AndyThompson #DataGovernance #EnterpriseAI #Snowflake #evolvConsulting #AIroi #DataQuality #SnowflakeCortex #Anthropic