The problem
A single wrong figure throws off the whole picture.
Before Dudley Land Company leases a parcel for oil and gas work, its landmen build a runsheet: a chronological record of every title document tied to that land, tracking each deed, lease, and assignment, and who owns what share. That ownership math, the fractional interest, decides who gets paid and how much, and a single wrong figure throws off the whole picture.
The runsheet goes to the client, who makes real leasing decisions on it. Reputation and money ride on it being right. Checking one by eye is slow, and a miscalculated interest is easy to miss in a long run of entries.
The solution
A ranked list of issues, critical items at the top.
After an AI training workshop with Benali, one of Dudley’s landmen put what he’d learned straight to work. Benali’s training walks a team through using Claude AI on their own real files, setting up projects and handing Claude the kind of documents they work with every day. Jeremy Jones tried it on title review the same afternoon. He set up a folder and asked Claude to review title work; when it asked for a completed runsheet, he exported one of the team’s stronger examples as a spreadsheet and dropped it in. Claude read what the landman had entered and handed back a ranked list of issues, critical items at the top, minor ones like spelling and numbering at the bottom.
At the top of that list sat a fractional-interest calculation that looked wrong, flagged as the first thing to check.
Jeremy did all of this on his first attempt, working from what he’d picked up in Benali’s training.
Time & impact
The whole review ran in about a minute and a half.
The whole review ran in about a minute and a half, way faster than reading every line by hand, and it surfaced the kind of error that’s easy to miss in a long run of entries.
The bigger result is who produced it: a Dudley landman, running his own review after training with Benali. The real value of the training is a team that can now run these reviews on its own and spend its judgment on the calls that actually need it.