How does JakBot know what to say? How does it answer questions about my business?
JakBot indexes your website content — pages, posts, and custom post types — by generating embedding vectors (a mathematical representation of your content’s meaning). When a visitor asks a question, JakBot converts their question into the same kind of vector and uses cosine similarity to find the most relevant content from your site. That content is then provided to the AI as context so it can generate an accurate, grounded answer. You can also configure tags, categories, and custom meta fields to enrich the embeddings and improve search quality for vague or category-level questions.