BER: The secret behind good AI-generated content

I’ve created hundreds of articles and other content forms with the assistance of artificial intelligence, more specifically LLM-techonology.

From using generative artificial intelligence in content generation workflows hundreds if not thousand of times, a pattern has emerged.

The best results seem the follow from a recipe something like this:

Starting from a RAG-enhanced base.

Enrich this base with knowledge from the LLM.

Finally reduce to a draft which I then always edit manually.

I liken this proces to cooking a good sauce. Starting with some good bases ingredients. Then enriching with spice. And reducing to get a concentrated taste. The doing the final touches before serving.