Using AI with Your Existing Content (What is RAG?)
Have you struggled with getting your organization's content (policies, procedures, guidelines, customer feedback, etc.) into your AI tools? The technical term for this is called RAG - Retrieval Augmented Generation. But you don't need to know what it's called, it simply means the AI tool processes the local content and mixes it with AI's broad knowledge. You get the benefit of vast LLMs with your specific organizational content factored in. In fact, it will weigh your local content more.
Photo: Rob O’Keefe
Why bring your content into AI?
Why do this? First, this gives you the benefit of private or organizational content into your results. Of course be careful that if you share that AI output you are not sharing sensitive or private information that has now been repackaged.
Second is to have your AI tool "learn" your or your organizations writing style, tone, and temperature. I put it in quotes as AI doesn't learn, but it identifies and mimics a pattern. It's important for content to be co-generated, enhanced by AI but not fully written by AI.
How to Do It?
There are many ways to accomplish this, the easiest is to drag a document into a Chat AI and it will process it. A more long standing way is to make a Project. Most platforms call this a Project (like ChatGPT, Claude, Grok), Microsoft calls it an Agent in Copilot, and Google has NotebookLM. This setup allows you to go back to a curated set of documents around a topic and define some custom prompting. You get more consistent and tailored responses, this is a big level up in your use of AI!
Scaling for the Enterprise?
At the enterprise level, organizations will learn to do this at scale. Using tools or custom LLM models to identify, curate, process, and make available the organizational content in AI ready form. Over more time, organizations will make this available externally via new AI interfaces such as Model Control Protocol (MCP), but more on that later.
Let me know in the comments or DM what topics you want me to cover about AI!
AI was only used in this post for spellcheck and a quick edit.