What an Automation Claim Owes You
Our homepage says 25 hours of drafting in 3 minutes. Any vendor can type that sentence. Here's the evidence standard we think a claim like that owes a solar engineer, and how ours is backed.
What an Automation Claim Owes You
The cover of our website says a 3.2 MW rooftop gets strung in about 3 minutes instead of 25 hours of hand drafting. Any vendor can type that sentence. Typing it costs nothing, and solar engineers know that, which is why most automation claims bounce off the people they're aimed at.
So instead of asking you to believe the sentence, this post lays out the standard we think any automation claim owes you, and shows how we try to meet it. Hold your other vendors to the same standard. Hold us to it too.
Claims are cheap. Receipts are not.
A drafting automation claim has three parts: the work, the time, and the correctness. Each part can be faked in a different way.
The work can be faked with a curated demo — a drawing picked because the tool handles it well, run once, recorded, edited. The time can be faked by starting the clock after setup. The correctness can be faked by simply not checking, because on a screenshot, a violating string looks exactly like a compliant one.
The fix for all three is the same: run the real thing where the reader can watch, and publish the checks alongside the result.
What we put behind the sentence
The stringing demo on our homepage is not a video. The page runs a live solve of the same rooftop, and the numbers on the cover — module counts, string counts, violations — come from that solve's output, not from marketing copy. If the solver produced a violation, the page would say so, because the page reads the data.
The correctness claim is stated as a count, not an adjective: zero NEC 690.7 violations out of the checked string set, with the voltage-window arithmetic available for an engineer to inspect. "Compliant" is an adjective. "0 of 12,000" is a number someone could catch us lying about.
And for the parts of the product that are not finished, our release room grades each claim in plain terms — live, or in build — with the evidence behind each grade published as it lands. The tool-authoring path in Leaf Automation Studio reads "in build" today, because it is. Writing that on our own marketing page cost us something, which is exactly why it is worth reading.
The questions to ask any vendor
If you're evaluating drafting automation — ours or anyone's — five questions separate a claim from a receipt:
- Can I run it on my drawing, not yours, before paying?
- Does the time quoted include setup, or start after it?
- Does the tool show the arithmetic behind each result, or just the result?
- When the tool is wrong, how do I find out — from the tool, or from the AHJ?
- Which claimed features are shipped, and which are roadmap wearing present tense?
A vendor with good answers will enjoy the questions. A vendor without them will change the subject to a demo video.
Why we publish it this way
Our customers are commercial and industrial solar engineers. They are the most claim-resistant audience in software, because they are professionally liable for believing wrong things. The only marketing that works on someone like that is evidence, so that's what we build: a live solve behind the cover, counts instead of adjectives, and a release room that says "in build" out loud.
We do the drafting. You do the engineering — including the engineering of deciding whether to believe us.
The live solve is on the homepage. The claim grades and their evidence are in the release room. If you'd rather test the claim than read about it, the trial is 14 days, on your own drawings.