S2 Ep5: Does a business have any business training an LLM?

About this episode

The episode discusses the trade-offs between training domain-specific LLMs and using foundational models, emphasizing the importance of Intelligent Document Processing and AI governance.

Enterprise LLMs: RAG vs Fine‑Tuning, IDP & Governance In this episode of the Mostly Unstructured podcast, Ed and Clay discuss whether it’s better to train a domain‑specific LLM or leverage foundational models like ChatGPT, Gemini and Claude. They explain the trade‑offs between fine‑tuning and retrieval‑augmented generation (RAG), and why Intelligent Document Processing (IDP) is vital for turning unstructured data into usable context. In this discussion, we cover: Why training your own LLM is risky and often unnecessary compared to adopting and building from a foundational model. How retrieval‑augmented generation (RAG) delivers more accurate results than simple fine‑tuning. The importance of Intelligent Document Processing (IDP) for ingesting unstructured data and building domain context. Real‑world lessons on AI governance, including the Air Canada bereavement‑policy chatbot case. Managing bias, hallucinations and toxicity in enterprise models. Measuring your return on AI investment. For those thrown by the excessive acronyms, let's define: LLM = Large Language Model RAG = Retrieval‑Augmented Generation IDP = Intelligent Document Processing. For more insights on enterprise…

People in this episode

Hosts: Ed, Clay

Topics covered

• Enterprise LLMs
• RAG vs Fine-Tuning
• Intelligent Document Processing
• AI governance
• Managing bias in AI

Keywords

• Large Language Model
• Retrieval-Augmented Generation
• Intelligent Document Processing
• AI governance
• enterprise AI
• data intelligence
• bias in AI
• hallucinations
• toxicity

Mentioned in this episode

Organizations: Air Canada, Keymark

Products: ChatGPT, Gemini, Claude

More episodes of The Mostly Unstructured Podcast

• S2 Ep9: Discoveries at the intelligence layer with Mike Askren · June 11, 2026 · 28 min
• S2 Ep8: Building Trust and Domain Agents with Matisha Ladiwala · June 4, 2026 · 19 min
• S2 Ep7: Are Data Lakes Drying Up? · May 13, 2026 · 38 min
• S2 Ep6: Agentic AI Needs Governance: Discussing Data Readiness, Audit Trails, and Trust · April 17, 2026 · 43 min
• S2 Ep4: A Selectively Deep Analysis of AI Analyst Predictions · February 17, 2026 · 30 min
• S2 Ep3: You know me I'm down with MCP · January 23, 2026 · 22 min