AI wants your data, but it can't handle the truth . . .

“We’ll just plug our data into ChatGPT.” Sure. Sounds easy. Rob thought so too, right up until he hit token limits, had to install Python, and discovered that even two megabytes of transcripts was too much for the world’s smartest models to handle. In this episode, Rob and Justin break down what happens when you try to use AI on your company’s internal data. Spoiler: it’s a lot more complicated than the vendors make it sound.

From semantic search to retrieval augmented generation (RAG), they unpack why the dream of “ask AI anything about your business” keeps falling short. You’ll hear why your Power BI model isn’t going anywhere, why structured data still needs old-school engines, and what it really takes to get value from your own information. Somewhere between a cautionary tale and a tech detective story, this one tells it like it is . . . unapologetically.

Run into the same situation? We’d love to hear about it! Give us a shout on LinkedIn and tell us how you overcame the limitations.