CTO says companies are missing out by not writing internal documents that AI can easily read

In a recent discussion, Laura Tacho, the CTO of developer productivity platform DX, highlighted a pressing issue that many companies face: the inadequacy of their internal documentation for artificial intelligence (AI) applications. Tacho emphasized that businesses should seriously consider restructuring their documents to enhance their compatibility with AI systems. Corporate documentation often consists of a chaotic mix of memos, slide presentations, and video training materials. According to Tacho, in this age of AI, it is essential for training materials to be primarily text-based. This format allows large language models (LLMs) to process information more efficiently. Developers, in particular, stand to gain significantly from this shift, as they can easily integrate text-based documentation into AI code editors like Cursor. Speaking on the "Pragmatic Engineer" podcast, Tacho explained that companies adopting an "AI-first" approach are already rethinking how they organize their internal documentation. She remarked, "That's been the biggest way that I've seen companies think about or already start trying to change the way that they architect their services." In a follow-up interview, Tacho elaborated on her documentation philosophy. She stressed that well-structured, text-based procedures benefit both humans and AI, stating, "The venn diagram is a circle." She pointed out that documentation often acts as a significant source of friction within organizations, and improving its quality can lead to substantial efficiency gains. Traditional documentation methods that rely heavily on visual elements, like screenshots, can be challenging for AI to interpret. Tacho advised that all images should include accompanying text descriptions to enhance accessibility. Drawing parallels with social media practices, she noted that platforms often use alternative texts and captions to support users who cannot rely solely on visual cues. Moreover, Tacho suggested that companies should centralize their documents to streamline both AI accessibility and human usability. She likened this process to "defragging"—a term used by developers to describe the organization of scattered data. "Documentation is made piecemeal, a little here, a little there," she explained, emphasizing the need for a cohesive structure to facilitate easier navigation. Tacho also mentioned technical aspects of documentation, such as ensuring semantic HTML markup to improve LLM readability. Some firms have already begun making these changes; for instance, Vercel's former VP of Developer Experience, Lee Robinson, shared on social media that the company is integrating cURL commands into their documentation, enhancing clarity for both users and AI. With the expectation that more businesses will follow suit, Tacho noted that engineers often waste over 30 minutes each week searching for information in poorly organized documentation. By improving document readability for AI, companies can automate and resolve these inefficiencies, transforming what she called a "very critical business problem" into an opportunity for enhanced productivity.