The Josh Bersin Company Microsoft Copilot Fine-Tuning With Galileo: Turn Copilot Into An HR Expert
18 snips
Nov 20, 2025 Discover the exciting potential of Microsoft Copilot's fine-tuning feature that can transform it into a tailored HR expert. Explore how embedding company-specific data enhances decision-making in management, leadership, and performance management. Learn about the advantages of fine-tuning over retrieval-augmented generation (RAG) and the practical HR use cases Galileo provides. With a focus on internal knowledge and trusted data, this innovation positions businesses to leverage AI for consistent and effective HR solutions.
AI Snips
Chapters
Transcript
Episode notes
Models Embed Data And Behavior
- Large language models embed training data and behavior in their neural network weights which shape answers and style.
- Josh Bersin explains that both content and tuning determine how models decide and generate responses.
Use RAG For Changing Content
- Use RAG to keep models up to date by feeding external documents for retrieval at query time.
- Josh Bersin recommends RAG for changing content like policies and pricing to avoid retraining.
Fine-Tuning Creates Company Voice
- Fine-tuning embeds specific corporate IP and voice into the core model so answers sound like your company.
- This produces responses that reflect firm policies, philosophies, and behaviors rather than generic outputs.