SMEs have today gained access to artificial intelligence through an as-a-service market consisting of language-model APIs, no-code tools, and verticalised applications, but they remain uncertain on what problems to tackle first, what tools to use, and how to do it in a sustainable way. This thesis seeks to leverage that accessibility to enable practical adoption. It begins by examining the historical barriers to entry for SMEs and describes how these have been dismantled, with inferences now roughly a thousand times cheaper than three years ago, but with contemporary entry barriers having moved away from technology to focus on the areas of skills and data.