Why Solar Engineers Keep Going Back to Manual

There's a pattern in solar engineering that software vendors don't like to talk about: an engineer evaluates a tool, uses it on a project or two, and quietly goes back to doing everything by hand in AutoCAD.

The vendor assumes stubbornness. That's not what's happening.

Three reasons. Every time.

1. "It couldn't do what I needed."

Most solar tools are built as end-to-end workflows. You enter at step one and the software walks you through to a finished output. On the demo, this looks incredible. The problem shows up on your third real project.

You need to route homeruns in a way the tool doesn't support. Your racking system isn't in the equipment library. The AHJ requires a labeling convention the software can't produce. And suddenly you're stuck — not because the engineering is hard, but because the tool has guardrails that won't let you deviate.

Tools that own the entire workflow are only as flexible as their least flexible feature. If step 4 can't handle your project, you can't skip to step 7. One incompatibility kills the entire tool for that entire project.

2. "I couldn't change what it produced."

The output was close but not right, and modifying it meant fighting the tool's internal data model instead of just editing a drawing. Close doesn't work in engineering — you can't submit a permit set that's 80% correct. So the engineer has two choices: manually fix the output, or do the whole thing from scratch. If fixing takes as long as redoing it, the tool provided no value. Worse, it consumed setup time you wouldn't have spent if you'd just opened AutoCAD.

3. "It was slower than doing it myself."

Setup time, learning curve, and workarounds for unsupported scenarios consumed more time than the tool saved. This is especially true for experienced engineers who are fast in AutoCAD — the bar for "faster than manual" is higher than most vendors assume. An experienced drafter can string a 500kW commercial rooftop manually in 3-5 hours. A tool that requires 2 hours of setup and produces output needing 2 hours of cleanup has saved nothing.

These aren't irrational complaints

They're the professional evaluation of someone whose livelihood depends on producing accurate construction documents on deadline. When a tool fails that evaluation, the rational response is to stop using it.

A different model

The alternative to an end-to-end tool isn't "no tool." It's a tool that fits into the engineer's workflow instead of replacing it.

Partial adoption has to work. Use stringing automation on this project, route homeruns by hand. No all-or-nothing.

Output has to be native and editable. Polylines, blocks, attributes — on the engineer's layers. Not proprietary objects that require the tool to view or edit.

Nothing is a black box. Every output is something the engineer can visually inspect, manually verify, and freely modify. Disagree with a string route? Grab the polyline and move it.

AutoCAD is a canvas, not a cage. Engineers who "go back to manual" are going back to the environment where they have complete control. That's not regression. That's risk management.

The tools that survive contact with real projects are the ones that respect this.

---

Leaf automates solar drafting inside AutoCAD — strings, homeruns, cable lengths, and tags. Everything it produces is a native CAD object you can modify, move, or delete. Start a 14-day free trial — no credit card required.