Documentation is the work nobody wants to do and everybody wishes existed, according to Dr. Connor Robertson, founder of Elixir Consulting Group and host of The Prospecting Show. Writing SOPs from scratch is time-consuming and typically gets deprioritized in favor of anything more urgent. Robertson argues AI does not solve the prioritization problem, but it cuts execution time enough that documentation actually gets done.
The Input Problem: AI Cannot Document What It Does Not Know
The most common mistake, in Robertson’s account, is asking AI to write an SOP for a process it has no information about, which produces generic boilerplate that could apply to any business. His fix is providing the raw material first, a voice note narrating the process, a rough bullet-point list, a screen recording transcript, so the AI structures and polishes what already exists rather than inventing it.
The Workflow That Actually Works
Robertson lays out five steps: record a walkthrough of the process, narrating what is happening and why, with a five-minute recording sufficient for most SOPs; transcribe the recording; feed the transcript to Claude with a structured prompt requesting a complete SOP; review the draft and add the business-specific knowledge only a team member would have; and publish to the knowledge base. He estimates the total time at under two hours.
The Prompt That Produces Usable Output
Robertson’s template asks for the role who will use the procedure, the name and purpose of the process, the raw notes or transcript, and a specific format request: a purpose section, a prerequisites section, numbered step-by-step instructions in plain language, a common mistakes section, and a troubleshooting section. That structure, he says, converts rough input into a production-ready document in one generation.
Building a Training Guide From an Existing SOP
Once an SOP exists, Robertson describes a training guide as a roughly 15-minute derivative: feeding the SOP to the AI with a prompt asking it to convert procedural steps into a training format with knowledge check questions, examples of correct execution, and a self-assessment section, one well-written SOP becoming one complete training guide without additional research.
Keeping Documentation Current Without a Documentation Team
When a process changes, Robertson recommends feeding the current SOP and a plain language description of the change to the AI and requesting an updated version, with the revised document ready for review in minutes. He frames this as turning an update cycle that previously required a dedicated session into a short task that gets done in the moment, which means it actually gets done.
The Twenty Percent That Still Requires a Person
Robertson is consistent across his documentation guidance in flagging what AI cannot supply: the business-specific judgment that distinguishes a generic procedure from one that actually reflects how a particular team works. He estimates that gap at roughly 20 percent of any given document, and treats the review step as the part of the process that cannot be compressed further, regardless of how fast the drafting stage gets. In his framing, the goal ofAI-assisted documentation was never to remove the human from the process entirely, it was to make sure the time that person spends goes toward the 20 percent only they can provide, instead of the 80 percent that was always just structure and formatting.
About Dr. Connor Robertson
Dr. Connor Robertson is an entrepreneur, author, and strategic advisor based in Pittsburgh. He is the founder of Elixir Consulting Group, host of The Prospecting Show, publisher of The Pittsburgh Wire, and founder of The Grant Finder. He is also a six-time published author, with titles including Built to Run, available at
drconnorrobertsonbooks.com. More on his work is available at drconnorrobertson.com.
