DataRobot unveils landman agent for tracking oil and gas lease rights

Turning scattered records into searchable lease data

Lease databases can show dates and other information entered by a team, but DataRobot says they may not capture the clauses and legal descriptions that determine which rights remain available. Those details can sit in an original lease, later assignments and underlying deeds recorded separately over decades.

The landman agent is designed to read scanned leases, deeds and assignments from a company’s document storage, including instruments dating back to the 1950s. It extracts fields such as ownership, primary term, start and expiration dates, Pugh clauses and legal descriptions. It also maps leases to the wells on them, connecting the lease record with drilling information.

This approach is aimed at questions that otherwise require a landman to locate and read the underlying documents, such as which leases expire in the next 60 days, which depths a Pugh clause has released and what drilling is needed to hold an existing lease.

Answers include the source instrument

According to DataRobot, users can ask questions in plain English and receive answers that cite the relevant source instrument. A landman can then check the cited page rather than rereading an entire file. The company says this shifts time from locating records to confirming information and acting on it.

The agent keeps corrections made by landmen. When a landman fixes an extracted field, that correction permanently overrides the extraction, while regression tests run across extraction runs so the change remains in place as new instruments are processed.

DataRobot says the agent does not decide which lease to extend, which depth to pursue or where to drill. It also does not communicate with mineral owners or approve title. Those judgments remain with the land team, while the agent supplies a connected record and the cited documents behind its answers.

A two-week proof of value covered 700 instruments

DataRobot says it built the agent with a leading domestic upstream oil and gas operator. During a two-week proof of value, the system ingested more than 700 scanned county instruments dating back to the 1950s and answered lease questions with the source document cited.

The company says the validation covered reading scans, extracting lease fields, answering in plain English with citations, retaining landman corrections, mapping leases to wells and regression testing across extraction runs. DataRobot describes this as a proof of value rather than a full deployment.

A broader deployment could add proactive alerts for upcoming expirations and open acreage, live document feeds, state regulatory production and permit data, maps of leases, units and wells, and support across multiple counties and states. DataRobot says which capabilities come first depends on a team’s acreage and priorities.

Designed for existing documents and controlled environments

The landman agent runs on the DataRobot Agent Workforce Platform. DataRobot says the platform can be deployed as managed SaaS, in a customer’s VPC, self-managed in its own environment or fully air-gapped. The agent reads documents and well data from local storage and works alongside an existing lease database, rather than requiring those documents to move to a new system.

DataRobot’s oil and gas team builds the agent on a customer’s documents and runs and governs it. The company says setup for a first county takes about two weeks using a sample of scanned instruments, with the customer providing the instruments and a landman to review the answers. The offering gives oil and gas teams a way to examine difficult-to-search lease records faster, while keeping source checking and decisions with their own land professionals.