Retrieval Augmented Generation, or RAG, is rapidly gaining popularity because it is the easiest way to customize generative AI solutions with your unique data. Unfortunately most people unfamiliar with large language models have a difficult time understanding it. Additionally, RAG solutions require a few weeks or months of upfront data work. As a result, you may struggle to get leadership support and funding for your RAG project, so you need to demonstrate potential impact before pitching it.

In Episode 19 I show you exactly how I do it with ChatGPT and a custom GPT in about an hour. I walk through a practical example where you, the analytics leader in an insurance company, are tasked with enhancing the customer support department. We guide you through the creation of a RAG demo that merges policy details from unstructured text documents with customer account information from databases. The value of RAG is instantly apparent when you share the results from a generic LLM with those supplemented with retrieved data. 

I document every step and share all of the prompts in Discover AI Opportunities with Generated Data.

These simple hacks are critical for analytics leaders, product managers and entrepreneurs trying to build support for their generative AI ideas. The best ideas don’t matter if you can’t get others to buy into them.

‍