Rosendo Torres - Diamondback Energy
Every upstream facilities project is different: a new oil gathering facility, a new well hookup, a brownfield upgrade on equipment that has been in service for decades. Execution documentation tailored to each site, including scope of work, quantity ledgers, daily progress and cost tracking, and turnover documentation, is what close-out KPIs and post-project analysis ultimately depend on. But building it by hand takes roughly a week per project, and on jobs that last three to four months, that effort has rarely penciled out. So the field has worked from general instructions, and the details between start and finish are captured informally, through notes, memory, and conversation, rather than tracked as the work happens. This presentation shares how an I&E supervisor is using AI to make site-specific execution documentation fast enough to be feasible on every project. With the pertinent project information in hand, AI-assisted generation cuts package build time from about a week to two days, and mid-project amendments take minutes instead of a rework. The first field application was an AI budget builder, built on historical labor costs, used to AFE a brownfield upgrade. The talk covers how the documents are generated and tailored to each site, what AI does well and where it still needs a human check, how field crews have received the packages, and lessons learned so far. The goal is simple: give every project the documentation it deserves, without the week of paperwork.