What happened when a gc stopped guessing on mep costs
1
He stopped padding numbers and started pulling them from the model.
MEP is where most estimators get burned, ductwork, conduit runs, and fixture counts buried
add comment
Answers
0
<p>He stopped padding numbers and started pulling them from the model.</p>
<p>MEP is where most estimators get burned — ductwork, conduit runs, and fixture counts buried in a model nobody has time to fully take off by hand. The usual fix is a rough multiplier per square foot, then hoping the bid holds. BidLight replaces that guess with a real takeoff.</p>
<p>Here's what changed once the GC brought MEP into BidLight instead of eyeballing it:</p>
<ul><li><strong>Exported the Revit model</strong> instead of manually counting fixtures, runs, and equipment off drawings.</li><li><strong>Let the two AI models classify BOQ line items</strong> by reading geometry and metadata, at roughly 86% accuracy — so mechanical, electrical, and plumbing items got sorted without a takeoff crew.</li><li><strong>Priced against a live database</strong> of about 30,000 items plus Craftsman, 1Build, and RSMeans, instead of a stale spreadsheet from the last job.</li><li><strong>Watched the estimate update in real time</strong> as the design changed, so a late duct reroute or panel swap didn't mean starting the MEP count over.</li><li><strong>Handed off a number he could defend</strong> in front of the owner, with line items tied back to the actual model instead of a gut-check percentage.</li></ul>
<p>The result isn't magic — it's just an estimate built from what's actually in the model instead of what usually gets built there. That's the whole point of running BidLight inside Revit: MEP stops being the line item everyone quietly pads and starts being one you can back up.</p>