I’ve been learning more about retrieval augmented generation (RAG) and how it can enhance AI responses and help AI-oriented businesses. Here are a few thoughts based on my understanding today:

• The better the data, the better the responses from AI. Data can be a moat that will keep competitors away. Data aggregated in a single source will be valuable because it will enhance the output from AI. Assuming that AI will aggregate disparate data, organize it, and provide a quality response isn’t a winning strategy as of today.
• Data structure matters for large data sets. AI struggles to make sense of large data sets. The data needs to be structured so AI can easily understand connections within it. And when I say data, I’m including text.
• AI will make it easy to build solutions. I see new niche AI apps launching daily. Competition is ramping up. To succeed long-term, entrepreneurs building AI solutions will need a proprietary data set or a unique way to distribute their solution to customers (e.g., brand credibility or a creative way to acquire customers)—ideally both. Absent these two things, having a sustainable edge over competitors will be difficult.

I’m still learning, but those are my thoughts right now.