Impact Land Services
Guides
Guide · Opinion

What “AI reads the lease” should mean.

Every AI title tool on the market now promises speed. Speed is not the question that decides whether you can trust the answer. The question is simpler: can you see the clause it came from?

Run a free EvalRead the spreadsheet guide →
Every finding cites its document and pageQuoted, never paraphrasedSummarizes. Never interprets.
Aerial view of section-line terrain in the Harney Basin, the kind of ground a land position is built from.
Every finding, sourced
Pugh clause — Addendum ¶19, page 3
What's being sold right now

Speed is the whole pitch.

In April 2026, one of the larger energy-data vendors announced a partnership with a title-automation platform, marketed on a specific number: up to 70 percent faster title timelines, with some materials citing 75 percent, document review pitched at 38 times faster than manual review, and a headline figure of $5 billion in royalty-calculation errors it says the workflow removes at the source. It rolled out first to Permian Basin customers.

We read the announcement the way we'd want a prospect to read ours: not for the headline number, for how the product explains itself. It talks about ownership calculations run automatically, a “title on demand” feature, and a risk score attached to each finding. Nowhere does it say how a user checks the score, or traces it back to the courthouse document or lease clause it came from.

That is not a knock on the number. 70 percent faster is a real claim from a real vendor, and it is probably true for the workflow it describes. It just answers the wrong question for the person who has to sign off on the result.

The one question that matters

Ask any AI title tool this, before you ask about speed:

“Show me the exact document and page this finding came from.”

If the answer is a confidence score or a risk rating with no clause behind it, you are being asked to sign off on a black box. Fast is worth nothing if you can't check the work, and the person whose name is on the payment is the one who has to check it.

How Impact's abstracts work

Every finding cites the language it came from.

Our AI reads the full lease package, every instrument, addendum and assignment as one file. Every term it flags, a royalty rate, a Pugh clause, a shut-in provision, a primary-term deadline, is attached to the specific paragraph and page it was pulled from, quoted, not paraphrased.

And the abstract summarizes. It does not interpret. It tells you what the lease says and where it says it; whether that supports your position is a question for your counsel, not our software. We are not a law firm, and this guide isn't legal advice either.

Hypothetical example, for illustration only

The same finding, two ways.

No client data here. A generic tract, built to show the difference between a score and a citation.

A risk-score output
TractSec 14, T21N R63W
Ownership riskMedium
Confidence82%

No clause, no page, no way to check it without re-reading the file yourself.

A cited finding
TractSec 14, T21N R63W
Pugh clauseUnit production holds only the unit
SourceAddendum ¶19, page 3

The clause is quoted where it lives in the document. Your counsel can check it in thirty seconds.

See the citations yourself.

Send us your first tracts. We run them through the system and hand back the findings with the clause attached, on your own position, not a demo of ours.

Run a free EvalSee the portal on sample data
What "AI reads the lease" should mean · Impact Land Services